Australia and New Zealand
Horizon Scanning Network
Continuous glucose monitoring devices
Horizon Scanning Report
NEW ZEALAND HEALTH TECHNOLOGY ASSESSMENT
DEPARTMENT OF PUBLIC HEALTH AND GENERAL PRACTICE
CHRISTCHURCH SCHOOL OF MEDICINE AND HEALTH SCIENCES
UNIVERSITY OF OTAGO
CHRISTCHURCH, NEW ZEALAND
© Commonwealth of Australia 2006
ISBN 1-74186-044-X
ISSN (Online)
This work is copyright. You may download, display, print and reproduce this material in
unaltered form only (retaining this notice) for your personal, non-commercial use or use
within your organisation. Apart from any use as permitted under the Copyright Act 1968,
all other rights are reserved. Requests and inquiries concerning reproduction and rights
should be addressed to Commonwealth Copyright Administration, Attorney General's
Department, Robert Garran Offices, National Circuit, Barton ACT 2600 or posted at
http://www.ag.gov.au/cca
Electronic copies of the Horizon Scanning Report can be obtained from:
http://www.horizonscanning.gov.au/
Disclaimer:
This report is based on information available at the time of research and cannot be
expected to cover any developments arising from subsequent improvements to health
technologies. This report is based on a limited literature search and is not a definitive
statement on the safety, effectiveness or cost-effectiveness of the health technology
covered. Before relying on the information in this report, users should carefully evaluate
its accuracy, currency, completeness and relevance for their purposes. This report is not
intended to be used as medical advice and it is not intended to be used to diagnose, treat,
cure or prevent any disease, nor should it be used for therapeutic purposes or as a
substitute for a health professional's advice. The Commonwealth does not accept any
liability for any injury, loss or damage incurred by use of or reliance on the information.
This Horizon Scanning Report was prepared by staff from the New Zealand Health
Technology Assessment group, Department of Public Health and General Practice,
Christchurch School of Medicine and Health Sciences, University of Otago, New
Zealand. http://nzhta.chmeds.ac.nz/ Mr Peter Day (Research Fellow) prepared the
report. The literature search strategy was developed and undertaken by Mrs Susan
Bidwell (Information Specialist Manager). Internal peer review was provided by
Dr Robert Weir (Director) and Mrs Susan Bidwell.
The Health Policy Advisory Committee on Technology (Health PACT) oversaw the
production of this Report. The Medical Services Advisory Committee (MSAC), the
Australian Health Ministers’ Advisory Council (AHMAC) and the New Zealand
Ministry of Health provide the funding for these Horizon Scanning activities.
Publication approval number: 3896
Table of Contents
Introduction.......................................................................................................................3
Background .......................................................................................................................4
Description of the technology ....................................................................................... 4
Development of the procedure.....................................................................................4
The procedure ..............................................................................................................5
Stage of development...................................................................................................5
Intended purpose..........................................................................................................8
Clinical need and burden of disease ............................................................................9
Burden of disease in Australia ...................................................................................9
Burden of disease in New Zealand ..........................................................................10
Summary..................................................................................................................10
Treatment Alternatives ..................................................................................................10
Existing comparators................................................................................................... 10
Clinical Outcomes ...........................................................................................................12
Diagnostic Accuracy ................................................................................................... 12
The diagnostic accuracy of CGMS® systems and the GlucoWatch® G2
Biographer evaluated in non-randomised controlled trials .......................................13
The diagnostic accuracy of the continuous glucose monitoring systems evaluated
in randomised controlled trials ..................................................................................13
The diagnostic accuracy of the STS™ Continuous Glucose Monitoring System ......13
The diagnostic accuracy of the Guardian® and the CGMS® systems........................14
Summary..................................................................................................................14
Effectiveness ............................................................................................................... 16
Diabetic control and the GlucoWatch® G2 Biographer.............................................16
Diabetic control and the CGMS® continuous glucose monitor .................................16
Diabetic control and the Guardian® CGMS continuous monitoring system .............17
Diabetic control and the STS™ Continuous Glucose Monitoring System................17
Summary....................................................................................................................21
Quality of life.............................................................................................................21
Summary..................................................................................................................22
Safety........................................................................................................................... 24
Hypoglycaemia and adverse events-The GlucoWatch® G2 Biographer ...................24
Hypoglycaemia and adverse events-The CGMS® continuous glucose monitoring
system ........................................................................................................................25
Hypoglycaemia and adverse outcomes-The STS™ Continuous Glucose Monitoring
System........................................................................................................................28
Summary..................................................................................................................28
Potential Cost Impact .....................................................................................................29
Cost Analysis .............................................................................................................. 29
Simple costings ..........................................................................................................30
Ethical Considerations....................................................................................................30
Informed Consent........................................................................................................ 30
Access Issues............................................................................................................... 31
Continuous glucose monitoring devices
1
Training and Accreditation............................................................................................31
Training ....................................................................................................................... 31
Clinical Guidelines...................................................................................................... 31
Sources of Further Information ....................................................................................32
Impact Summary ............................................................................................................33
Conclusions......................................................................................................................34
Appendix: Levels of Evidence........................................................................................38
Search Strategy ...............................................................................................................39
Limitations of the Assessment........................................................................................40
References........................................................................................................................41
Tables
Table 1: Specifications of continuous glucose monitoring systems approved or pending
approval for clinical use ......................................................................................7
Table 2 Diagnostic accuracy Dexcom STS™ Continuous Glucose Monitoring
System ...............................................................................................................14
Table 3 Diagnostic accuracy of the CGMS® and Guardian® Monitoring Systems ......15
Table 4 Diabetic control from the GlucoWatch® G2 Biographer ..................................18
Table 5 Diabetic control from the CGMS® ....................................................................19
Table 5 (continued) Diabetic control from the CGMS® ...............................................20
Table 6 Diabetic control from the Guardian® CGMS ....................................................20
Table 7 Diabetic control in the DexCom STS™ Continuous Glucose Monitoring
System ...............................................................................................................21
Table 8 Quality of life ....................................................................................................23
Table 9 Hypoglycaemic alarm and adverse outcomes with the GlucoWatch® G2
Biographer .........................................................................................................25
Table 10 Hypoglycaemic alarm and adverse outcomes and the CGMS® ........................26
Table 10 (continued) Hypoglycaemia and adverse outcomes and the CGMS® ............27
Table 11 Hypoglycaemic alarm and adverse events and the STS™ system ...................28
Table 12 Designations of levels of evidence ...................................................................38
Table 13 Levels of evidence for assessing diagnostic accuracya .....................................39
Table 14 Literature sources utilised in assessment ..........................................................39
Table 15 Search terms utilised.........................................................................................40
2
Continuous glucose monitoring devices
Introduction
The New Zealand Health Technology Assessment Unit, Department of Public Health
and General Practice, Christchurch School of Medicine and Health Sciences, University
of Otago, on behalf of the Medical Services Advisory Committee (MSAC) and the New
Zealand Ministry of Health, has undertaken an Horizon Scanning Report to provide
advice to the Health Policy Advisory Committee on Technology (Health PACT) on the
state of play of the introduction and use of continuous glucose monitoring (CGM)
devices for patients with diabetes mellitus.
Continuous glucose monitoring (CGM) devices provide in depth information about
fluctuations in glucose levels throughout the day and facilitate the prevention of hypoand hyperglycaemia in patients with diabetes mellitus. A range of technologies are
currently being researched for non-invasive and minimally invasive continuous glucose
monitoring systems. Only impedance spectroscopy (the application of electromagnetic
radiation through the skin to the blood vessels) and interstitial fluid (ISF) technologies
are currently approved for clinical use.
Typically these devices consist of a small monitor that reads and displays glucose values
in real-time or retrospectively, a glucose sensor which is attached subcutaneously to the
abdomen or externally to the arm, and a transmitter to relay information about glucose
concentrations between the sensor and monitor. These monitoring devices supplement
but do not replace standard blood glucose self-monitoring (SMBG) practices and can be
used in home settings. Seven CGM devices are currently U.S. Food and Drug
Administration (FDA) approved for clinical use on a prescription basis in the U.S. or CE
(Conformité Européene or European Conformity) marked for use in Europe. Other
devices are pending FDA approval. In Australia and New Zealand, the CGMS®
Continuous Glucose Monitoring System (Medtronic MiniMed, Northridge, CA) and its
second generation replacement CGMS® System Gold™ are currently available but only
for approved physician-supervised use for inpatients and are pending TGA approval in
Australia.
This Horizon Scanning Report is intended for the use of health planners and policy
makers. It provides an updated assessment of the current state of development of
continuous glucose monitoring devices in general, their present use, the potential future
application of the technology, and its likely impact on the Australian and New Zealand
health care systems. This Horizon Scanning Report is an update statement based on the
latest available evidence derived from Randomised Controlled Trials (RCTs) on the
safety, effectiveness, cost-effectiveness and ethical considerations associated with
continuous glucose monitoring devices for diabetic patients.
Continuous glucose monitoring devices
3
Background
Description of the technology
Development of the procedure
The Diabetes Control and Companion Trial (DCCT) (1993) and the U.K. Prospective
Diabetes Study (UKPDS) (1998) have shown that tight glycaemic control and intensified
insulin therapy reduces long-term complications associated with diabetes mellitus.
However, intensified insulin therapy also increases the risk of severe hypoglycaemic
events. Frequent and accurate blood glucose monitoring is necessary if the disease is to
be managed optimally. Standard self-testing techniques with finger pricks and home
glucose meters can be painful, messy, and inconvenient and have poor patient
compliance. Due to the importance of achieving euglycaemia in diabetes management
and limitations with conventional SMBG methods there has been considerable
investment and development in CGM technologies with more than 100 companies
currently involved in the research (Sieg et al, 2005).
Glucose monitoring has in the past relied on blood glucose levels. Newer technologies
using non-invasive and minimally invasive glucose sensing technologies focus on the
interaction of electromagnetic radiation with tissue and the harvesting of interstitial fluid
(ISF) across the skin.
Truly non-invasive techniques involve tissue irradiation, the analysis of the absorbed and
scattered radiation, the processing of this information, and the measurement of the
glucose in the dermal tissue. This includes optical methods such as near-infrared, Raman
spectroscopy, polarimetry, light scattering, and photoacoustic spectroscopy. The
structural and physiological properties of the skin present difficulties in the reliability of
these optical methods for glucose monitoring sensing. At present there are no large-scale
clinical studies supporting the efficacy, portability, and affordability of these
technologies. Therefore at present minimally invasive technologies are offering the
greatest potential for practical continuous glucose monitoring devices for clinical use
(Bui et al, 2005; Sieg et al, 2005).
Minimally invasive continuous glucose monitoring techniques sample and monitor
glucose concentrations in the ISF and not the blood. These techniques are considered to
be minimally invasive because they do not puncture any blood vessels but rather they
bring a sensor into contact with ISF by inserting a sensor subcutaneously (into the
abdominal wall, wrist or arm) to measure ISF in situ (e.g. direct sensor implantation,
CGMS®) or by extracting this fluid to an external sensor by various mechanisms which
compromise the skin barrier (e.g. reverse iontopheresis, GlucoWatch® G2 Biographer)
(Klonoff, 2005a).
A wide range of continuous glucose monitoring technologies are currently being
researched. At this stage only non-invasive impedance spectroscopy (the application of
electromagnetic radiation through the skin to the blood vessels) and minimally invasive
technologies (interstitial fluid (ISF) measurement in situ or extraction through the skin)
are approved for clinical use.
4
Continuous glucose monitoring devices
The procedure
Typical CGM systems consist of:•
•
•
a small monitor (usually a pager sized recording device clipped to the belt) that
reads and displays glucose values in real-time or retrospectively
a glucose sensor (usually a glucose-oxidase based needle-like sensor) which is
implanted subcutaneously in the abdomen (or externally to the wrist, arm or
forearm)
a transmitter to relay information about glucose concentrations between the
sensor and monitor (via a cable or wirelessly).
Some devices such as the CGMS® System Gold™ have a Com-station, a communication
device for downloading sensor data to a computer. Each device has specific attractive
features such as sensor placement options, longer sensor life, alarms for out-of range
values, real time readings (a feature in more recent devices), and varying degrees in
invasiveness. Each device undergoes a warm-up period of 1-2 hours, a device specific
calibration process of between 1 and 4 times per day, and each device’s sensor provides a
blood glucose reading every 1-10 minutes for up to 72 hours and up to 3-months for
newer non-invasive technology. The blood glucose information is available to the
patient and clinician either in real time or retrospectively and many models have alarms
that trigger if glucose levels fall outside of preset euglycaemic ranges (Table 1.)
(Klonoff, 2005a).
Stage of development
Seven CGMs have been approved by the U.S. Food and Drug Administration (FDA) for
clinical use in the U.S. or carry the CE marking for clinical use in Europe. The basic
specifications of these devices are outlined in Table 1.
The devices are the:• Continuous Glucose Monitoring System Gold (CGMS® System Gold™;
Medtronic MiniMed, Northridge, CA)
• GlucoWatch® G2 Biographer (Cygnus Inc, Redwood City, CA)
• Guardian® Telemetered Glucose Monitoring System (Medtronic MiniMed,
Northridge, CA)
• Guardian® RT Continuous Glucose Monitoring System (Medtronic MiniMed,
Northridge, CA)
• GlucoDay® (Menarini Diagnostics, Florence, Italy)
• Pendra (Pendragon Medical, Zurich, Switzerland)
• STS™ Continuous Glucose Monitoring System (Dexcom, San Diego, CA)
One further device, the FreeStyle Navigator™ Continuous Glucose Monitor (Abbott
Laboratories, Alameda, CA), is pending FDA approval.
Minimally invasive CGM systems are available in Australia which measure interstitial
blood glucose via an indwelling sensor in the subcutaneous tissue of the abdomen or
buttocks. Only the CGMS® Continuous Monitoring System (Medtronic MiniMed,
Northridge, CA) and its successor the CGMS® System Gold™ are currently available in
Australia and New Zealand but are only available in approved hospitals, diabetes clinics
Continuous glucose monitoring devices
5
and research centres for physician-supervised use in the management of specialised
diabetes cases (Cameron and Ambler, 2004). In Australia it is currently not approved by
the TGA and clinicians must apply to the TGA for use in each specialised patient
(personal communication, Medtronic Australasia). Non-invasive CGM systems such as
the GlucoWatch® G2 Biographer are not available in Australia or New Zealand. Some
individuals have accessed this CGM technology via overseas contacts and it is likely that
increasing patient demand will see this technology introduced.
6
Continuous glucose monitoring devices
Table 1: Specifications of continuous glucose monitoring systems approved or pending approval for clinical use
Product
FDA
Approved/
CE
marked
Year first
approved
or marked
Sensor
type
Sensor
mechanism
Sensor
location
Sensor
warmup
(h)
Calibrations
per lifetime
of sensor
Sensor
lifespan
(h)
Frequency
of
testing
(min)
Time of
blood
glucose data
display
Alarm
Continuous Glucose
Monitoring System
(CGMS)
Gold
(Medronic MiniMed)
Yes/Yes
1999
Minimally
invasive
Enzyme-tipped
catheter
Subcutaneous
abdomen
2
12
72
5
Retrospective
No
GlucoWatch G2
Biographer (GW2B)
(Cygnus Inc)
Yes/ Yes
2001/2002
Minimally
invasive
Reverse
iontophoresis
External on
arm or
forearm
2
1
13
10
Real time
Yes
Guardian Telemetered
Glucose
Monitoring System
(Medtronic MiniMed)
Yes/Yes
2004
Minimally
invasive
Enzyme-tipped
catheter
Subcutaneous
arm
2
12
72
5
Retrospective
Yes
GlucoDay
(Menarini
Diagnostics)
No/Yes
2001
Minimally
invasive
Microdialysis
Subcutaneous
abdomen
0
1
48
3
Real time or
retrospective
Yes
Pendra
(Pendragon Medical)
No/Yes
2004
Noninvasive
Impedance
spectroscopy
External on
wrist
1
20
3 months
1
Real time
Yes
FreeStyle Navigator
Continuous
Glucose Monitor
(Abbott Laboratories)
No/No
—
Minimally
invasive
Enzyme-tipped
catheter
Subcutaneous
arm or
abdomen
1
1
72
1
Real time
Yes
Guardian RT
Continuous Glucose
Monitoring System
(Medtronic MiniMed)
Yes/No
2005
Minimally
invasive
Enzyme-tipped
catheter
Subcutaneous
abdomen
2
6
72
5
Real time
Yes
STS System
(DexCom)
Yes/No
2005
Minimally
invasive
Enzyme-tipped
catheter
Subcutaneous
abdomen
2
6
72
5
Real time
Yes
Adapted from Klonoff (2005a)
Continuous glucose monitoring devices
7
Intended purpose
Testing blood glucose levels is essential for children, adolescents and adults who
suffer from Type-1 and Type-2 diabetes requiring treatment with insulin and for
people taking oral hypoglycaemic agents. A primary treatment goal in insulin
dependent diabetes is tight glucose control to minimise complications associated with
diabetes. Insulin dependent patients must monitor their blood glucose to ensure that
suitable levels of insulin are circulating. To ensure appropriate glucose control
patients may require intensive therapy to administer insulin by injection at least three
or more times during the day, or have insulin delivered by an external or implanted
pump by continuous insulin infusion. There are varying recommendations regarding
the method, technique, timing, and frequency of self-monitoring blood glucose
(SMBG) in non-insulin dependent Type-2 diabetes patients (Bergenstal and Gavin,
2005).
SMBG with the finger prick test and glucose meter adds valuable information that
complements glycohaemoglobin (HbA1c) testing in achieving optimal glycaemic
control. Conventional point-in-time blood testing provides discrete and highly
accurate data about current glucose levels. Given that HbA1c monitoring is a timeaveraged result it has limitations as a marker for glycaemic control. This is because it
does not represent “real time” monitoring for patients and physicians and excursions
in glycaemic control such as post-prandial hyperglycaemia or severe hypoglycaemia
may be missed (Bergenstal and Gavin, 2005). The main limitations with SMBG are
that this method is invasive, messy and painful, and inconvenient which can affect
testing compliance, especially among paediatric patients.
Minimally invasive continuous glucose monitoring systems obtain multiple readings
over time but are not as accurate as SMBG testing and for people taking oral
hypoglycaemic agents. CGM devices are used to obtain in depth information about
changes in blood glucose levels throughout the day to facilitate optimal glycaemic
control for Type-1 and Type-2 diabetes patients requiring insulin therapy. Patients
and clinicians have the ability to view real time blood glucose values, to
retrospectively review recent blood glucose values, review trend graphs for glucose
values, and receive alerts/alarms for impending hypo- or hyperglycaemia. These
devices are particularly useful for night time monitoring particularly in preventing
episodes of hypoglycaemia unawareness. Continuous glucose monitoring devices
supplement but do not replace standard blood glucose self-monitoring (SMBG)
practices and can be used in home settings.
CGM is also being used in the management of patients with conditions associated
with diabetes. The CGMS® monitoring device has been used to monitor children with
diseases associated with hypoglycaemia including hyperinsulinism (Conrad et al,
2004), and glycogen storage (Hershkovitz et al, 2001), hyperglycaemia in pregnancy
(Buhling et al, 2004; Kerssen et al, 2004), glucose monitoring of neonates in neonatal
intensive care (Beardsall et al, 2005; Javid et al, 2005), and to assist with continuous
positive airway pressure (CPAP) interventions for obstructive sleep apnea (Babu et al,
2005; Czupryniak et al, 2005). Patients with cystic fibrosis related diabetes have
benefited from CGM (Jefferies et al, 2005), as have pregnant women where strict
glycaemic control can reduce perinatal complications (Kerssen et al 2005).
8
Continuous glucose monitoring devices
CGM has been utilised as an adjunct monitoring system to assess glycaemic control
using continuous subcutaneous insulin infusion (CSII) compared with multiple daily
injection (MDI) for both Type-1 and Type-2 diabetes patients (Hirsch et al, 2005;
Wainstein et al, 2005). Prototype systems for CGM in closed-loop systems with a
combination of a continuous glucose monitor, a control algorithm, and an insulin
pump have been developed. The main two approaches use either a minimally
invasive subcutaneous system for glucose monitoring and insulin delivery or
intravenous glucose sampling and intraperitoneal insulin delivery (Hovorka, 2006).
Medtronic MiniMed (Northridge, CA) have particularly been involved in closed-loop
projects utilising both the CGMS® and Guardian® monitoring systems and other
groups have also utilised Medtronic MiniMed technology. Although feasibility
studies into closed-loop systems have been conducted the reliability and accuracy of
the glucose monitors remains the main limitation to commercially viable closed-loop
systems (Hovorka, 2006).
Clinical need and burden of disease
Type-1 or juvenile diabetes occurs where sufferers produce no (or very little) insulin
due to the auto-immune destruction of the insulin producing beta cells of the pancreas.
Type-1 diabetes represents approximately 10-15 per cent of all diabetic patients,
however 98 per cent of childhood diabetes is Type-1. Type-2 diabetes occurs where
sufferers still produce insulin but production is impeded and is characterised by
reduced levels of insulin or insulin resistance and represents approximately 85-90 per
cent of diabetic sufferers, most of whom are over the age of 40 years. Gestational
diabetes is a temporary form of diabetes, which occurs during pregnancy in 3-8 per
cent of females not previously diagnosed with diabetes (AIHW, 2002).
It is recommended that blood glucose self-monitoring be performed by checking
blood glucose levels at least 3 times per day using the finger prick test and glucose
meter (Bergenstal and Gavin, 2005). This method is invasive, messy, and can be
painful and difficult to perform.
Burden of disease in Australia
An estimate of the number of patients who could benefit from continuous glucose
monitoring was based on estimates of diabetes prevalence. The Australian Institute of
Health and Welfare (AIHW) 2001 National Health Survey estimates, based on selfreported information, that 95,000 (0.5%) of Australians have type 1 diabetes and that
a further 900,000 (7%) of adults aged 25 years and over had type 2 diabetes based on
1999/2000 data (AIHW, 2004).
Estimates for the age-standardised prevalence of Type-1 diabetes for 1999-2000 was
298 per 100,000 or approximately 37,000 individuals over the age of 25 years
(AIHW, 2002). Since 1999, the National Diabetes Register (NDR) has collected
information on the number of new users of insulin. There were 4,548 new cases of
Type-1 diabetes, aged 0-39 years, for the years 1999-2001, 50 per cent of these cases
were children aged 0-14 years. The most recent data on the incidence of childhood
diabetes in Australia for the years 2000-2001 indicate an incidence rate of 20.3 and
18.9 per 100,000 for males and females, respectively (AIHW, 2003). Of the 21,346
new insulin users registered on the NDR for the years 1999-2001, 12,167 (57%)
suffered from Type-2 diabetes. The majority (90%) of these patients were aged over
Continuous glucose monitoring devices
9
35 years (AIHW, 2003). The prevalence of Type-2 diabetes is increasing and has been
associated with obesity, poor nutrition and physical inactivity.
Over the period 2001–2003 there were 20,908 diabetes-related deaths registered
(5.4% of all deaths) for people aged 25 years or over in Australia. Diabetes was
recorded as the underlying cause of death in 9,772 of these cases, representing 2.5%
of all deaths registered during the period (AIHW, 2005a).
The total Australian health system expenditure on diabetes in 2000–01 was estimated
to be around $784 million, or 1.7% of allocatable recurrent health expenditure.
Diabetes was ranked fifteenth out of around 200 disease groups compared. An
estimated $204 million was spent by the Australian Government on people with
diabetes on antidiabetic drugs and diabetes testing reagents. Although only 10% of
the 4.6 million prescriptions for antidiabetic drugs in 2000–01 were for insulin, these
accounted for 60% of expenditure on antidiabetic drugs. Average health expenditure
on diabetes in 2000–01 was $1,469 per known (self-reported) case of diabetes, or $42
per Australian (AIHW, 2005b).
Burden of disease in New Zealand
The New Zealand population of adults aged 15 years and over in 2002/03 was
estimated at 3,124,690 (Statistics New Zealand, 2005). Self-reported diabetes
information from the 2002/03 New Zealand Health Survey indicated that the diabetes
prevalence rate in adults aged 15 years and over was 4.2%, of which 85-90% were for
type 2 diabetes (Ministry of Health, 2004). Based on these prevalence estimates
131,237 adults aged 15 years and over have diabetes where an estimated 111,551 to
118,113 persons have Type-2 diabetes. The estimated incidence of Type-1 juvenile
diabetes was 25.8 cases per 100,000 persons aged up to 19 years in 2001. Over the
period from 1999-2001 there were 2,324 deaths registered (2.8% of all deaths) where
diabetes was recorded as the underlying cause (Ministry of Health, 2005).
Summary
There is huge potential for the use of CGM within the diabetic population of Australia
and New Zealand. Based on self-reported data available approximately one million
Australians could possibly benefit from CGM devices. In New Zealand based on selfreported information, more than 131,000 adults over 15 years could potentially benefit
from CGM devices.
Treatment Alternatives
Existing comparators
Material on existing comparators was sourced from the ANZHSN report on the
GlucoWatch® G2 Biographer (Australia and New Zealand Horizon Scanning
Network, 2005). The optimal method for assessing long-term glycaemic control is the
measurement of glycosylated haemoglobin. Haemoglobin combines with blood
glucose to form glycosylated haemoglobin or HbA1c. When plasma glucose is
10
Continuous glucose monitoring devices
consistently elevated there is a corresponding increase in levels of HbA1c stored in
erythrocytes. Due to the 120 day life span of erythrocytes, the levels of HbA1c will
reflect the glycaemic history of the patient over the past 2-3 months. HbA1c levels
determined by high-performance liquid chromatography (HPLC) are the standard
reference for glycosylated haemoglobin measurements.
Levels of HbA1c should reflect to a certain extent glucose levels determined by selfmonitoring of blood glucose (SMBG). When measured by HPLC, a HbA1c level of 6
per cent approximates a plasma glucose level of 6.6 mmol/L or 120 mg/dL. A 1 per
cent rise in the HbA1c level equates to a 1.7 mmol/L or 30 mg/dL increase in the
mean glucose level (Braunwald et al 2001; FDA 2002). The normal average value for
preprandial glucose is <5.5 mmol/L, with an ideal range of 4.4-6.7 mmol/L
(Braunwald et al 2001). Patient action should be taken for values <4.4 or >7.8
mmol/L. Similarly the normal average value for bedtime glucose is <6.1 mmol/L,
with an ideal range of 5.5-7.8 mmol/L, with action required if values are <5.5 or >8.8
mmol/L. Hypoglycaemia and hyperglycaemia may be defined as plasma glucose
levels of <2.5 mmol/L and 28 mmol/L, respectively. However these levels may vary
with symptoms and physiologic responses (Braunwald et al 2001).
The current gold standard for SMBG for use by the patient in the home is the glucose
meter, which is a small, portable battery operated device. There are currently more
than 25 different brands of commercially available glucose meters, including AccuChek® Advantage® (Roche Diagnostic), One Touch® (LifeScan Inc) and Accutrend®
DM (Boehringer Mannheim). SMBG is recommended for all people with diabetes,
but especially for those treated with insulin. It is recommended that patients with
Type-1 diabetes test glucose levels three or more times per day. SMBG plans may
recommend testing glucose levels before all meals, two hours after meals and before
retiring for the night. To test glucose levels patients should wash hands thoroughly to
remove any trace of glucose and reduce risk of infection, prick the fingertip with a
lancet and hold the finger until a large droplet of blood forms. The droplet of blood is
placed onto a test strip, which is then inserted into the glucose meter. The test strip is
coated with glucose oxidase, which then converts any glucose present in the blood to
hydrogen peroxide. A dye impregnated into the test strip combines with the hydrogen
peroxide and, when placed into the glucose meter, will reflect light according to the
amount of glucose present. Higher glucose concentrations will reflect less light.
Glucose meters should be calibrated regularly using a standard glucose solution
(FDA, 2002).
All portable blood glucose meters measure the amount of glucose in whole blood.
Glucose levels in plasma are generally 10-15 per cent higher than glucose
measurements in whole blood. The results are displayed on a digital readout
approximately 1-2 minutes after the test strip is placed into the meter. Glucose meters
can detect glucose over the range 0-34 mmol/L. Many SMBG meters now give results
as "plasma equivalent", using a built in algorithm, allowing comparison of home
glucose measurements to those determined from plasma by HPLC (FDA, 2002).
Continuous glucose monitoring devices are only intended to supplement information
acquired from conventional SMBG. They are not intended to replace SMBG by the
patient in the home with a glucose meter.
Continuous glucose monitoring devices
11
Clinical Outcomes
The GlucoWatch® G2 Biographer and the original CGMS® Continuous Glucose
Monitoring System and its second-generation replacement the CGMS® System Gold™
(the sensor and software were modified in 2002) were the first clinically approved and
commercially available devices and have been studied more extensively in the
medical literature than other devices. The RCT studies in this report mainly utilised
the original CGMS®, the GlucoWatch® G2 Biographer and in one RCT, the
Guardian® RT Continuous Glucose Monitoring System and in another the STS™
Continuous Glucose Monitoring System. Other Horizon Scanning reports have
recently been published for the GlucoWatch® G2 Biographer (Australian and New
Zealand Horizon Scanning Network, 2005; AETSA, 2005a) and minimally invasive
blood glucose monitoring systems (CGMS® and GlucoDay®) (AETSA, 2005b).
Diagnostic Accuracy
Standard glucose monitoring measures blood glucose levels but the development of
minimally invasive glucose sensing techniques has focused on measuring glucose in
the ISF. Accurate measurement of glucose is dependent upon a strong correlation
between glucose in the ISF and the blood. Glucose enters the ISF from the blood and
is removed again by uptake into the surrounding cells. Under normal physiological
conditions the process of glucose exchange between the plasma and the interstitial
space means that changes in the glucose concentration in the ISF are strongly
correlated with those in the blood (Seig et al, 2005). However, changes in blood flow
and metabolic rate can affect glucose concentrations in the ISF with resulting changes
in the ISF not always reflected in the blood. There can be a lag time that can vary
between 0 and 45 minutes. Also glucose concentrations in the ISF can also precede
the blood level and/or change at a greater rate and can complicate the interpretation of
the glucose measurement in the ISF (Seig et al 2005, Kulcu et al, 2003; Aussedat et al,
2000). Models have shown how that a rapid change in blood glucose (e.g. from
insulin injection) will be reflected in the ISF with a delay.
The relationship between glucose concentrations in the ISF and blood is also
important as all current CGM techniques require calibration with a blood sample
obtained by the conventional method prior to sensor monitoring. This calibration
value is used in subsequent monitoring. Inaccuracies in calibration will result in
inaccurate glucose measurements, hypo- and hyperglycaemic alarm settings and
insulin therapy adjustments (Choleau et al, 2002). Deviations from the glucose
reference values in the blood may be due not only to technical limitations but also
may be due to physiological factors. Further research into the complex relationship
between blood and the interstitial compartment is required in order to achieve
standardised and objective performance evaluation for CGM devices and accuracy for
the optimisation of metabolic control (Sieg et al, 2005).
Point to point glucose comparisons between SMBG and CGM monitoring for
determining accuracy are limited in their analysis of glucose trends. Other methods
have been developed, such as the continuous grid-error analysis (CG-EGA) which
utilises point- and rate-error analysis to capture the point presentation and the
direction and rate of blood glucose fluctuations for hypoglycaemic, euglycaemic and
hyperglycaemic ranges (Kovatchev et al, 2004). The accuracy of the TheraSense
12
Continuous glucose monitoring devices
FreeStyle™ Navigator Continuous Glucose Monitor (Abbott Laboratories, Alameda,
CA) was assessed using grid-error analysis and the failure to detect hypoglycaemia at
blood glucose extremes (73.5% accuracy) was the most common error. The device
was reported to be very accurate in periods of euglycaemia (99%) and hyperglycaemia
(95.4%).
The diagnostic accuracy of CGMS® systems and the GlucoWatch® G2 Biographer
evaluated in non-randomised controlled trials
The accuracy of the first U.S. FDA approved CGM devices, the original Medtronic
MiniMed CGMS® and its modified successor the CGMS® System Gold™ and the
Cygnus GlucoWatch® G2 Biographer (GWB) have been more rigorously assessed in
clinical studies than other devices. The Diabetes Research in Children Network
(DirecNet), a multi-center collaborative study group, has produced a number of nonrandomised studies on the accuracy of both of these devices in children and
adolescents with Type-1 diabetes. These studies found that both sensors were less
accurate in measuring glucose levels during hypoglycaemia and that gender, ethnicity,
Body Mass Index (BMI), or age (3-18 years) had no effect on the function of either
sensor. The CGMS® sensors were equally accurate in measuring glucose levels on
each of the three days of wear while the GWB was less accurate in the last 5-hours of
wear. In comparing day and night-time accuracy the GWB showed no differences in
accuracy whilst the CGMS® was less accurate with lower readings at night
(Buckingham et al, 2005).
The diagnostic accuracy of the continuous glucose monitoring systems evaluated in
randomised controlled trials
Diagnostic accuracy outcomes for CGM devices were reported in four RCT studies.
These compared glucose level measurements derived from CGM sensors compared to
those obtained from SMBG. These studies were graded as level 3b evidence
according to the levels of evidence for grading diagnostic accuracy (refer to evidence
grading hierarchy in the Appendix, Table 13).
The diagnostic accuracy of the STS™ Continuous Glucose Monitoring System
One RCT study by Garg et al (2006) (level 3b evidence) reported on the accuracy of
real-time sensor values for the Dexcom STS™ Continuous Glucose Monitoring
System compared with SMBG values. Overall 95.4% of paired glucose values were
within Clarke error grid A and B zones with a correlation coefficient of 0.88. There
were mean and median absolute differences but sensor values were within prespecified accuracy limits with SMBG glucose values and no systematic bias was
detected (Table 2).
Continuous glucose monitoring devices
13
Table 2 Diagnostic accuracy Dexcom STS™ Continuous Glucose Monitoring System
Study
Population
Outcomes
47 Type-1 and Type-2
diabetic patients wearing
DexCom STS for CGM
(with display data provided
only during 72-hr periods 2
and 3)
Prospective real-time sensor values were
compared with SMBG values, 95.4% of 6,767
paired glucose values within Clarke error grid A
and B zones. Correlation coefficient was 0.88.
Features
Garg et al
(2006), USA
Level 3b
evidence
RCT
44 Type-1 and Type-2
diabetic control patients
wearing DexCom STS for
CGM (with data not
provided during any 72hour period)
Mean absolute difference
21.2%
Median absolute difference
15.9%
No systematic bias detected at pre-specified
glucose levels (50, 80, 100, 150, 200 mg/dL)
For the 91 subjects mean
age 44± 13 years and 75
patients with Type-1
diabetes and 16 with Type2 diabetes
Study conducted in a
clinic/home environment
The diagnostic accuracy of the Guardian® and the CGMS® systems
Three RCT studies reported diagnostic accuracy (level 3b evidence) in the Medtronic
MiniMed Guardian® and the Medtronic MiniMed CGMS® glucose sensors. The
study by Bode et al (2004) assessed an earlier version of the Guardian® continuous
monitoring system that did not provide real time sensor glucose values. A later model
has been developed with real-time functionality, the Guardian® RT Subcutaneous
Glucose Monitoring System (SGMS). The accuracy of the Guardian® was measured
using the absolute relative error (ARE) where a lower ARE indicates greater
equivalency between sensor and home meter blood glucose readings.
The study by Tansey et al (2005) evaluated a modified CGMS® sensor. The overall
median relative absolute difference (RAD) in readings was 12% and CGMS®
accuracy was greater during periods of hyperglycaemia than hypoglycaemia and did
not vary significantly by the number of calibrations entered.
The study by Fiallo-Scharer et al (2005) compared glucose monitoring with the
CGMS® sensor and eight-point SMBG over 3-days. This study was part of the same
trial for the studies by Chase et al (2005) and The Diabetes Research in Children
Network (DirecNet) Study group (2006). Both monitoring regimes gave similar mean
glucose profiles in terms of target ranges and associations with HbA1c but only 10%
of subjects completed 3-days of eight-point home glucose testing. The CGMS®
sensor tended to overestimate postprandial excursions (p<0.001) and underestimate
mean overnight glucose levels (p<0.001) (Table 3).
Summary
In these studies the CGMS® sensors were less accurate than home sensors and less
accurate during periods of hypoglycaemia compared to SMBG. The Guardian® was
also less accurate during periods of hypoglycaemia with nearly half of all alerts being
false alerts and consistently lower readings than concurrent home SMBG readings.
14
Continuous glucose monitoring devices
With the STS® monitoring device pairs of sensor and SMBG glucose values were well
correlated and differences were within pre-specified accuracy limits with SMBG
glucose values and no systematic bias was detected.
Table 3 Diagnostic accuracy of the CGMS® and Guardian® Monitoring Systems
Study
Population
Outcomes
35 type-1 diabetic patients
with the Guardian CGMS
with 2 alert sensors
switched off followed by 2
sensors alerts switched on
in the alert group
Hypoglycemia alert for the Guardian
Mean age 42.2 ± 10.3
years, range (24-68)
63% sensitivity, 97% specificity, 19% false alerts
Features
Bode et al
(2004), USA
Level 3b
evidence
RCT,
accuracy
assessed
using single
sample
comparative
study of all
alert and nonalert data
Tansey et al
(2005), USA
Level 3b
evidence
As a part of
an RCT,
a crossclassification
study
Fiallo-Scharer
et al (2005),
USA
Level 3b
evidence
Cross-over
RCT
36 type-1 diabetic controls
with all 4 Guardian CGMS
sensors alerts switched off
Mean age 45.8 ± 12.2
years, range (24 -71)
Study conducted in the
home environment
≤ 70 mg/dL
67% sensitivity, 90% specificity, 46% false alerts
Hyperglycemia alert for the Guardian
≥ 250 mg/dL
Mean (median) absolute relative error between
home SMBG and the Guardian was 21.3%
(17,3%).
Average readings for Guardian were 12.8 mg/dL
below the concurrent home SMBG readings
Two hundred children
enrolled in an outpatient
setting, 191 children
included in analysis
1,899 CGMS-reference pairs, median pairs per
subject 10. Mean calibrations per 24h of CGMS
was 5.3.
191 Type-1 diabetic
patients wearing the CGMS
over 2-3 days compared
with reference home
glucose meter
Difference (mg/dL)
Mean age 12.5±2.8 years,
Study conducted in the
home environment
Two hundred children
enrolled
161 Type-1 diabetic
patients wearing CGMS
over 3 days
Mean age 12.4 years,
range (7-17)
161 Type-1 diabetic
controls with 8-point testing
using meters over 3 days
Mean age 12.4 years,
range (7-17)
Study conducted in the
home environment
Median, 25%/75% percentiles
-4 (-24,13)
Absolute diff (mg/dL ) 18 (8, 38)
Relative difference
a
RAD
b
ISO criteria met
-3% (-15%, 9%)
12% (6%, 23%)*
72% (69%, 74%)**
(mean and 95%CI)
* p<0.001; ** p<0.002
Patients used CGMS an average of 70 h over 3
days
Average readings per subject
859
Eight-point testing per subject
19
Eight-point
Mean glucose (mg/dL) 188 ± 41
(overnight)
199
In target range***
50%
Above target range
47%
Below target range
3%
Postprandial excursions
(Mean mg/dL)
17
Correlation HbA1c and
mean glucose (mg/dL) 0.40**
CGMS
183± 37*
174**
49%
46%
5%
63**
0.39**
* p<0.009; ** p<0.001; *** 61-180 mg/dL
a
Median relative absolute difference (difference divided by reference value*100) b International Organisation for
Standardisation reference glucose values for ≤75 mg/dL, CGMS value ± 15 mg/dL for glucose values for 75 mg/dL, CGMS
value ± 20 mg/dL
Continuous glucose monitoring devices
15
Effectiveness
The optimal method for assessing overall and long-term glycaemic control and the
effectiveness of glucose monitoring with CGM devices or SMBG in the management
of diabetes is the measurement of glycosylated haemoglobin (HbA1c). When plasma
glucose is consistently elevated there is a corresponding increase in levels of HbA1c
stored in erythrocytes. Due to the 120 day life span of erythrocytes, the levels of
HbA1c will reflect the glycaemic history of the patient over the past 2-3 months.
Levels of HbA1c should reflect to a certain extent glucose levels measured by SMBG
and those measured by CGM.
Diabetic control and the GlucoWatch® G2 Biographer
Two RCTs (level II evidence) evaluated the effectiveness of the GlucoWatch® G2
Biographer blood glucose monitoring system on HbA1c levels in paediatric patients
with Type-1 diabetes. A retrospective review by physicians of blood glucose profiles
and a real-time alarm function for impending hypo- and hyperglycaemia from the
GlucoWatch® G2 Biographer was used to adjust therapy regimes. Safety and
effectiveness outcomes for the study by Chase et al (2003) was sourced from the
ANZHSN report on the GlucoWatch® G2 Biographer (Australia and New Zealand
Horizon Scanning Network, 2005) (Table 4).
An RCT by Chase et al (2003) (level II evidence) monitored the HbA1c levels of 20
paediatric patients wearing a GlucoWatch® G2 Biographer for three months and 20
patients using SMBG. This study reported that the HbA1c levels of the GlucoWatch®
G2 Biographer patients were significantly lower (8.4%), which may indicate
improved glycaemic control, than those patients using standard care alone (9.0%,
p<0.05), at the end of the three month intervention period.
Another RCT by Chase et al (2005) (level II evidence) compared the GlucoWatch®
G2 Biographer with SMBG in 200 children and adolescents with Type-1 diabetes
(also see studies by The Diabetes Research in Children Network (DirecNet) Study
group 2006 and Fiallo-Scharer et al 2005 as part of same trial). There were no
significant decreases in HbA1c at six months (p=0.15) in either group. There was no
significant difference between groups in the proportion of HbA1c decreases of
≥ 0.5%.
Diabetic control and the CGMS® continuous glucose monitor
Four RCTs (two with level II evidence and two with poorer quality study design with
level III-1 evidence) evaluated the effectiveness of the CGMS® continuous glucose
monitor plus SMBG on HbA1c levels in paediatric and adult patients with Type-1
diabetes. Therapy regimes were adjusted by physicians and patients based on
retrospective review of glycaemic profiles from the CGMS in these four trials (Table
5).
A smaller pseudo-randomised RCT with eleven children with Type-1 diabetes by
Chase et al (2001) (level III-1 evidence) compared the CGMS® with SMBG. After
the first month all five subjects in the CGMS® group decreased their HbA1c levels by
at least 0.2%. At 3-months there was no significant decrease in HbA1c in the CGMS®
group (p=0.07) or the control group (p>0.05). This study was limited by the very
small numbers of children included and did not compare monitoring groups.
16
Continuous glucose monitoring devices
The RCT by Ludvigsson and Hanas (2003) (level II evidence) compared subjects with
CGMS® plus SMBG in an open trial arm with subjects using the CGMS® blinded to
CGMS® monitoring profiles in 27 paediatric subjects with Type-1 diabetes. At 3months HbA1c levels had significantly decreased in the open trial arm patients
(p=0.013) but not the patients blinded to CGMS® profiles and there was a statistically
significant difference in HbA1c levels between groups (p=0.011).
The study by Tannenburg et al (2004) (level II evidence) compared CGMS® in 128
adults with Type-1 diabetes with SMBG. At 3-months both the intervention and
control groups of subjects showed statistically significant improvements in HbA1c
levels (p<0.001). For between-group comparison there were similar improvements in
HbA1c levels at 3-months. The CGMS® group had a significantly reduced duration of
hypoglycaemia at the end of the study compared to baseline (p=0.009).
The pseudo-randomised RCT by Chico et al (2003) (level III-1 evidence) compared
the CGMS® in 75 Type-1 adult diabetes patients with SMBG. At 3-months both the
intervention and control subjects showed improvements in HbA1c levels that were
statistically significant (p<0.01). A between-group comparison of improvements in
HbA1c levels at 3-months was not done. A subgroup analysis of patients who
underwent continuous subcutaneous insulin infusion (CSII) also showed greater
decreases in HbA1c, regardless of monitoring method, (p<0.01) than those who
underwent multiple insulin infusion (MII) (p<0.05).
Diabetic control and the Guardian® CGMS continuous monitoring system
One RCT, Bode et al (2004) (level II evidence), compared the Guardian® CGMS
continuous monitoring system (sensor alert switched on) with the Guardian® CGMS
(sensor alert switched off) in 71 adult patients with Type-1 diabetes. This Guardian®
CGMS model did not provide real time sensor glucose values but provided alarms that
activated when user excursions go below or above predetermined thresholds of hypoand hyperglycaemia. The sensor alert group had significantly (p=0.03) reduced
median hypoglycaemia excursion periods compared to the control group without
sensor alert and there was no differences in the frequency of hypoglycaemia
excursions (Table 6).
Diabetic control and the STS™ Continuous Glucose Monitoring System
One RCT by Garg et al (2006) (level II evidence) evaluated the STS™ continuous
glucose monitoring system plus SMBG where 47 Type-1 and Type-2 diabetes patients
over two 72-hour periods were provided with real-time glucose values, trend
information, and received alerts and alarms compared with 44 controls who were not
provided with any of this information . Patients in the display group spent
significantly less time as hypoglycaemic (especially during night time) and
hyperglycaemic, and significantly more time within blood glucose level targets
(p<0.001) (Table 7).
Continuous glucose monitoring devices
17
®
Table 4 Diabetic control from the GlucoWatch G2 Biographer
Study
Population
Outcomes
Chase et al
(2003), USA
20 Type-1 diabetic patients
wearing GlucoWatch G2
Biographer:
Patients used biographer an average of 3.5
times per week
Level II
evidence
Mean age 11.9 ± 3.1 years,
range (7-16)
3.6% readings
>300 mg/dL (16.7 mmol/L)
RCT*
20 Type-1 diabetic controls
with SMBG:
15.5% readings
<70 mg/dL (3.9 mmol/L)
Mean age 11.9 ± 3.3 years,
range (7 -17)
HbA1c% (median)
Features
*Some
authors
affiliated with
Cygnus Inc,
Redwood, CA
Study conducted in the
home environment
Total readings
GlucoWatch®
G2 Biographer
Baseline
8.9
3 months
8.4
6 months
8.3
9 months
8.4
11,925
Control
8.6
9.0
8.5
8.6
At 3 months HbA1c was significantly lower in
GlucoWatch® G2 Biographer compared to
control groups (p< 0.05), Wilcoxon Rank Sum
test.
Chase et al
(2005), USA
Level II
evidence
RCT
99 Type-1 diabetic patients
wearing GlucoWatch G2
Biographer for CGM over 6
months:
Patients with the biographer used an average of
1.2 ± 0.7 sensors per week with at least 8-hours
of use, and 16% subjects at least 2.0 uses of
8≥ hours/week.
Mean age 12.3± 2.7 years,
range (7-17)
Usage declined throughout the study from mean
2.1 times/week in 1st month to 1.5 times/week in
th
the 6 month. Reasons for declining usage:
101 Type-1 diabetic
controls having SMBG over
6 months:
Mean age 12.7± 2.9 years,
range (7 -17)
Study conducted in the
home environment
18
Usual care
GW2B
Mean HbA1c (6 months)
7.9%
8.1%
95% CI mean reduction (-0.4% to 0.1%; p=0.15)
Decrease on HbA1c ≥ 0.5% 21%
*p=0.21
28%*
Continuous glucose monitoring devices
Table 5 Diabetic control from the CGMS®
Study
Population
Outcomes
5 Type-1 diabetic patients
with the CGMS:
Patients using CGMS used 6 3-day sensors
within a 30 day period. Both groups selfmonitored at least 4-times daily.
Features
Chase et al
(2001), USA
Level III-1
evidence
RCT
Mean age 14.8 ± 2.2 years,
range (10-17)
Sensor readings
421.4 (mean) 86.7 (sd)
Sensor hours
35.1 (mean)
6 Type-1 diabetic controls
with SMBG:
Mean age 12.6 ± 0.6 years,
range (11 -13)
Study conducted in the
home environment
7.2 (sd)
Controls
CGMS
Baseline
9.0± 1.2
10.0± 0.7
1-month
8.8± 0.4
9.5± 0.9*
3-month
8.4± 0.2
8.8± 0.3
Mean HbA1c levels %
Mean decrease in HbA1c %
1-month
Tanenberg et
al (2004),
USA
Level II
evidence
RCT*
*Sponsored
by Medtronic
MiniMed,
Northridge
CA
Ludvigsson
and Hanas
(2003), USA
Level II
evidence
Cross-over
RCT*
*Supported
by grant from
Medtronic
MinMed,
Northridge
CA
0.2± 0.2
0.36± 0.07
p=0.37
p<0.01
3-month
0.62± 0.44
*p<0.01
p>0.05
1.04± 0.43
p=0.07
51 Type-1 & 2 diabetic
patients with the CGMS:
Patients using CGMS used self-monitoring 4.0±
1.7 times daily and SMBG was 3.9± 1.6 times
daily.
Mean age 44.0 ± 10.2
years
Mean HbA1c levels (%)
128 patients enrolled
58 (54 analysed) Type-1 &
2 diabetic controls with
SMBG:
Mean age 44.5 ± 12.6
years
Study conducted in the
home environment
SMBG
Baseline
9.0± 1.0
CGMS
9.1± 1.1
2-month
8.3± 0.9*
8.3± 0.9*
3-month
8.3± 0.9
8.3± 0.9
*p<0.001 improvement from baseline for each
group
Similar improvement from baseline between
groups p=0.95
32 patients enrolled and
randomised, 27
participated.
Patients using CGMS used self-monitoring 4.0±
1.7 times daily and SMBG was 3.9± 1.6 times
daily.
Mean age 12.5 ± 3.3 years,
range 5-19 years
Average sensor life 2.1± 1.0 days
16 type-1 diabetic patients
with the CGMS in an open
trial arm and insulin
therapy adjusted
accordingly by team/patient
16 type-1 diabetic patients
with the CGMS in a blinded
study arm with insulin
therapy adjusted based
solely on SMBG
Study conducted in the
home environment
Continuous glucose monitoring devices
298 sensor profiles in both arms
CGMS
CGMS
Open
blinded
(SMBG)
Mean HbA1c levels (%)
Baseline
7.70
7.75
3-month
7.31*
7.65**
*p=0.013 improvement from baseline
**p=0.011 difference between the two arms
19
Table 5 (continued) Diabetic control from the CGMS®
Study
Population
Outcomes
Chico et al
(2003), Spain
105 patients enrolled
Mean number of readings with CGMS 816±179
and glucose meter 19±6 in 75 Type-1 diabetic
patients.
Level III-1
evidence
Mean age 36.5± 12 years
Features
RCT
40 Type-1 diabetic patients
with the CGMS:
35 Type-1 diabetic
controls SMBG
Mean age 41± 10 years
Plus CGMS arm compared
with 30 additional Type 2
diabetic patients for
asymptomatic
hypoglycemias:
Mean age 58± 11 years
Study conducted in the
home environment
SMBG
CGMS
Mean HbA1c levels (%)
Baseline
8.0± 1.4
8.3± 1.6
3-month
7.5± 0.8*
7.5± 1.2*
*p<0.01 improvement from baseline
Subgroup analysis of patients also showed
decreases in HbA1c in both groups with
continuous subcutaneous insulin infusion (CSII)
(p<0.01) and multiple insulin infusion (MII)
(p<0.05).
Table 6 Diabetic control from the Guardian® CGMS
Study
Population
Outcomes
Patients used 322 Guardian CGMS sensors
over 699 cumulative days
Level II
evidence
35 type-1 diabetic patients
with the Guardian CGMS
with 2 alert sensors
switched off followed by 2
sensors alerts switched on
in the alert group:
RCT
Mean age 42.2 ± 10.3
years, range (24-68)
Features
Bode et al
(2004), USA
36 type-1 diabetic controls
with all 4 Guardian CGMS
sensors alerts switched off:
Mean age 45.8 ± 12.2
years, range (24 -71)
Study conducted in the
home environment
Average home readings per day
6.4
Average calibration values per day
3.9
Average paired readings 7.0 per day the Alert
group and 5.8 for the Control group
Non-Alert
Alert
Guardian
Guardian
(control)
Change from baseline
(period 2-1)
Hypoglycemia excursions (minutes)
(Median change of time) -4.5
-27.8*
Hypoglycemia excursions (number)
(frequency)
+0.5
+0.03
Hyperglycemia excursions (minutes)
(Median change of time) +9.9
-9.6
Hyperglycemia excursions (number)
(frequency)
+0.2
+0.1
* p=0.03
20
Continuous glucose monitoring devices
Table 7 Diabetic control in the DexCom STS™ Continuous Glucose Monitoring System
Study
Population
Outcomes
47 Type-1 and Type-2
diabetic patients wearing
STS for CGM (with display
data provided only during
72-hr periods 2 and 3)
Compared with the control group, the display
group spent
44 Type-1 and Type-2
diabetic control patients
wearing STS for CGM (with
data not provided during
any 72-hour period)
26% more time on target (81-140 mg/dL)*
*p<0.001
Features
Garg et al
(2006), USA
Level II
evidence
RCT
For the 91 subjects mean
age 44± 13 years. 75
patients with Type-1
diabetes and 16 with Type2 diabetes.
21% less time as hypoglycaemic (< 55 mg/dL)*
23% less time hyperglycaemic (≥ 240 mg/dL)*
Nocturnal hypoglycemia (time~hours)
-38% reduction (<55 mg/dL)
-33% reduction (<55-80 mg/dL)
Study conducted in a
clinic/home environment
Summary
CGM devices were of limited effectiveness in assisting with glycaemic control in the
short term (up to 3-months) as measured by HbA1c levels. The measurements of
glucose levels with CGM devices were used to make adjustments and generally
resulted in small improvements in HbA1c levels in monitored patients. However, the
quality of the evidence in these studies was limited by a lack of between-group
comparison with SMBG and the very select (mainly children and adolescents) and
small patient samples assessed in the studies.
Quality of life
Three RCT studies (two level II and one with poorer quality study design with level
III-1 evidence) reported quality of life outcomes derived from a range of validated and
reliable psychometric questionnaires.
The RCT by Chase et al (2003) (level II evidence) reported on the quality of life of
paediatric patients with Type-1 diabetes wearing the GlucoWatch® G2 Biographer
compared to those with SMBG and found no difference between the two groups at the
end of the three month intervention (Table 8).
An RCT by The Diabetes Research in Children Network (DirecNet) Study group
(2006) (level II evidence) compared the GlucoWatch® G2 Biographer with SMBG in
200 children and adolescents with Type-1 diabetes (also see studies by Chase et al
2005 and Fiallo-Scharer et al 2005 as part of same trial). The validity and reliability
of the Fear of Hypoglycaemia and Quality of Life questionnaires had been assessed.
There were no significant differences or changes between parents and youths and
glucose monitoring groups between baseline and at 6-months for the Diabetes Worry
Scale. There was some indication of a small deterioration in diabetes selfmanagement over the period as Management Profile scores significantly decreased for
youths and parents for both monitoring groups (p<0.001) but there were no
Continuous glucose monitoring devices
21
differences between groups. The Paediatric Quality of Life scale (PedsQL) showed
that parent’s scores were significantly higher than youth’s and both increased slightly
over the period indicating deterioration in quality of life but there was no significant
differences between monitoring groups.
One small RCT with eleven children with Type-1 diabetes by Chase et al (2001)
(level III-1 evidence) compared the CGMS® with SMBG. The validity and reliability
of the Fear of Hypoglycaemia and Quality of Life questionnaires were assessed.
There were no significant differences between the two groups of subjects at baseline,
1-month and 3-months in the Fear of Hypoglycemia and Quality of Life survey
questionnaire results. This study was limited by the very small number of children
included.
Summary
There was a lack of clear evidence indicating significant improvements in quality of
life for patients using CGM devices compared to SMBG as assessed by various
psychometric measures.
22
Continuous glucose monitoring devices
Table 8 Quality of life
Study
Features
Chase et al
(2003), USA
Level II
evidence
Population
Outcomes
20 Type-1 diabetic patients
wearing GlucoWatch G2
Biographer:
Fear of hypoglycaemia scores a at 3 months
Mean age 11.9 ± 3.1 years,
range (7-16)
RCT*
*Some
authors
affiliated with
Cygnus Inc,
Redwood, CA
The Diabetes
Research in
Children
Network
(DirecNet)
Study group
(2006), USA
Level II
evidence
20 Type-1 diabetic controls
with SMBG:
Mean age 11.9 ± 3.3 years,
range (7 -17)
Study conducted in the
home environment
RCT
GlucoWatch® G2 Biographer
81.3 ± 11.7
Control
79.8 ± 15.5
Psychological measuresa at 6-months for
parents and youths for GW2B and SC groups
Mean age 12.3± 2.7 years,
range (7-17)
Diabetes Self-Management Profile
Scores decreased over 6-months for both youths
and parents for both groups (p<0.001) but no
differences between groups.
Diabetes Worry Scale
No significant changes or differences between
groups over the 6-months
PedsQL Diabetes Module
Parents scores significantly higher than youth
score.
Scores increased over the 6-months among both
groups (p=0.16).
No differences between GlucoWatch and SMBG
groups.
101 Type-1 diabetic
controls having SMBG over
6 months:
Mean age 12.7± 2.9 years,
range (7 -17)
Study conducted in the
home environment
Level II
evidence
DCCT b Quality of Life score a at 3 months
99 Type-1 diabetic patients
wearing GlucoWatch G2
Biographer over 6 months:
RCT
Chase et al
(2001), USA
®
GlucoWatch G2 Biographer
59.0 ± 14.3
Control
56.4 ± 9.6
5 Type-1 diabetic patients
with the CGMS:
Mean age 14.8 ± 2.2 years,
range (10-17)
6 Type-1 diabetic controls
with SMBG:
Mean age 12.6 ± 0.6 years,
range (11 -13)
No significant differences between the two
groups of subjects in results for fear of
Hypoglycemia and Quality of life survey
questionnaires for baseline, 1-month and 3months after.
Fear of Hypoglycemia score decreased slightly
in CGMS group from 61.8 to 56.6 at 3-months
(p>0.05).
Study conducted in the
home environment
a
b
questionnaires are reliable and validated, DCCT = Diabetes control and complications trial, RCT=randomised
controlled trial
Continuous glucose monitoring devices
23
Safety
Outcomes considered in the studies assessed were the CGM alarm function in the
detection of the frequency and duration of hypoglycaemia, and symptomatic and nonsymptomatic hypoglycaemia during day and night time periods, adverse outcomes
related to device malfunction and monitor-related adverse reactions such as skin
rashes.
Hypoglycaemia and adverse events-The GlucoWatch® G2 Biographer
Two RCT studies (level II evidence) reported on hypoglycaemia and other adverse
outcomes for the GlucoWatch® G2 Biographer device compared to a control group
using SMBG (Table 9).
The study by Chase et al (2003) (level II evidence), reported that hypoglycaemic
events were detected significantly more often in the GlucoWatch® G2 Biographer
intervention group than the control group (p<0.005), and those patients wearing the
GlucoWatch® G2 Biographer intermittently were able to detect hypoglycaemic events
more easily even when not wearing the device (p<0.03), which may reflect increased
patient awareness of nocturnal hypoglycaemia resulting from the experience of
wearing the GlucoWatch® G2 Biographer device. However, GlucoWatch® G2
Biographer was not able to detect all hypoglycaemic events, missing approximately
21 per cent of those events that actually occurred and were confirmed by conventional
finger-prick testing. In addition, the proportion of readings below 70 mg/dL (3.9
mmol/L), and therefore in the hypoglycaemic range, increased over the course of the
three month intervention, from 14.2 to 16.5 per cent of readings (p<0.002). The
increase in hypoglycaemic readings may be due to a more aggressive approach to
glycaemic management by the patients over the course of the intervention.
The RCT by Chase et al (2005) (level II evidence) compared the GlucoWatch® G2
Biographer with SMBG in children and adolescents with Type-1 diabetes (also see
studies by The Diabetes Research in Children Network (DirecNet) Study group 2006
and Fiallo-Scharer et al 2005 as part of same trial). Sensor use declined during the
study from a mean 2.1 times/week in the 1st month to 1.5 times/week in the 6th
month. The reasons provided for declining usage were skin reactions with
approximately half of patients experiencing moderate or acute skin problems and
sensor technical difficulties. The addition of the GlucoWatch® G2 Biographer to
standard blood glucose monitoring did not decrease severe hypoglycaemic events in
that group of patients. These were at least three times greater in the GWB group than
the SMBG group.
24
Continuous glucose monitoring devices
Table 9 Hypoglycaemic alarm and adverse outcomes with the GlucoWatch® G2 Biographer
Study
Features
Chase et al
(2003), USA
Level II
evidence
Population
Outcomes
20 Type-1 diabetic patients
wearing GlucoWatch G2
Biographer:
Mean age 11.9 ± 3.1 years,
range (7-16)
Hypoglycaemia
(≤70 mg/dL or 3.9 mmol/L)
Detected more frequently in GlucoWatch® G2
Biographer compared to control group
χ2, p<0.0005
20 Type-1 diabetic controls
with SMBG:
There were 42 episodes of hypoglycaemia,
78.6% of these were registered by GlucoWatch®
G2 Biographer
RCT*
*Some
authors
affiliated with
Cygnus Inc,
Redwood, CA
Chase et al
(2005), USA
Level II
evidence
RCT
Mean age 11.9 ± 3.3 years,
range (7 -17)
Study conducted in the
home environment
Detected more frequently in GlucoWatch® G2
Biographer at night, compared to control group
even when NOT wearing device
χ2, p<0.03
Percent of GlucoWatch® G2 Biographer
readings < 70mg/dL during intervention
phase
Month 1
Month 2
14.2%
16.6%
Month 3
16.5%
99 Type-1 diabetic patients
wearing GlucoWatch G2
Biographer for CGM over 6
months:
Usage declined throughout the study from mean
2.1 times/week in 1st month to 1.5 times/week in
th
the 6 month. Reasons for declining usage from
questionnaire:
Mean age 12.3± 2.7 years,
range (7-17)
Skin irritation (76%), frequent skips (56%),
excessive alarms (47%), inaccurate readings
(44%).
101 Type-1 diabetic
controls having SMBG over
6 months:
Mean age 12.7± 2.9 years,
range (7 -17)
Study conducted in the
home environment
Severe hypoglycaemia
events
**p=0.10
Usual care
2%
GW2B
7%**
Skin reactions from GW2B use: all subjects
reported at least once over 6-months. One
subject severe and 48 subjects (48%) moderate.
At the 6-month follow-up 54 (55%) subjects had
acute changes (mild 36%, moderate 19%,
severe 0%). 50 (51%) subjects had non-acute
changes such as scabbing, dry skin, hypohyperpigmentation or scarring.
Hypoglycaemia and adverse events-The CGMS® continuous glucose monitoring
system
Four RCT studies (two level II and two with poorer quality study design with level
III-1 evidence) reported on hypoglycaemia and adverse outcomes for the CGMS®
continuous glucose monitoring system compared to a control group with SMBG. The
CGMS® sensor monitoring was always in addition to SMBG in the intervention group
(Table 10).
Chase et al (2001) (level III-1 evidence) compared the CGMS® with SMBG in
children with Type-1 diabetes. There were significantly more mean hypoglycaemic
episodes (<60 mg/dL) detected per subject in the first month in the CGMS® group,
p=0.001. The majority of night time episodes were asymptomatic in the CGMS®
group. No severe adverse events occurred during the study which may indicate that
CGMS® monitoring occurred without an increase in the risk of severe hypoglycaemia.
Continuous glucose monitoring devices
25
The RCT by Tannenburg et al (2004) (level II evidence) compared the CGMS® with
SMBG. At 3-months both groups of subjects showed no statistically significant
differences in the frequency and duration of hypoglycaemia. There were two severe
hypoglycaemic events and five monitor-related adverse reactions in the CGMS®
group compared to one severe hypoglycaemic event in the SMBG group.
Ludvigsson and Hanas (2003) (level II evidence) compared subjects with CGMS®
plus SMBG in an open trial arm with subjects using the CGMS® plus SMBG in a
blinded study arm. At 3-months there was no statistical significant difference in the
frequency of hypoglycaemic episodes between treatment arms. However, the duration
of night-time hypoglycaemic episodes was over twice the duration of day time
episodes. There was one case of severe hypoglycaemia in each study arm. Only onethird of subjects had a 3-day curve for sensor functionality data.
The study by Chico et al (2003) (level III-1 evidence) compared the CGMS® with
SMBG in Type-1 and Type-2 diabetes patients. At 3-months Type-1 diabetes patients
showed more asymptomatic hypoglycaemic episodes than Type-2 patients with night
time episodes being more prevalent than during the day. Over half the patients
monitored by CGMS® had unrecognised hypoglycaemic episodes detected.
Table 10 Hypoglycaemic alarm and adverse outcomes and the CGMS®
Study
Population
Outcomes
Features
Chase et al
(2001), USA
Level III-1
evidence
RCT
5 Type-1 diabetic patients
with the CGMS:
Controls
Mean hypoglycaemic
Mean age 14.8 ± 2.2 years,
range (10-17)
episodes (<60 mg/dL)
6 Type-1 diabetic controls
with SMBG:
Night time episodes
6.7± 1.1
CGMS
12.8± 1.6*
per subject 1st month
Mean age 12.6 ± 0.6 years,
range (11 -13)
Asymptomatic
Study conducted in the
home environment
No seizures, episode requiring help, or
unconscious episodes in either group during
study
Symptomatic
17
4
3
* p=0.001
26
Continuous glucose monitoring devices
Table 10 (continued) Hypoglycaemia and adverse outcomes and the CGMS®
Study
Population
Outcomes
Features
Tanenberg et
al (2004),
USA
Level II
evidence
RCT*
*Sponsored
by Medtronic
MiniMed,
Northridge
California.
Ludvigsson et
al (2003),
USA
Level II
evidence
Cross-over
RCT*
*Supported
by grant from
Medtronic
MinMed,
Northridge
CA
SMBG
128 patients enrolled
CGMS
51 Type-1 & 2 diabetic
patients with the CGMS:
Hypoglycaemia (n)*
Mean age 44.0 ± 10.2
years
At 3-months
58 (54 analysed) Type-1 &
2 diabetic controls with
SMBG:
Daytime
1.2± 1.0
1.1± 0.8
Total 24h
1.7± 1.2
1.4± 1.1
Events per week
2.3± 2.3
1.9± 1.6
0.5± 0.5
0.4± 0.4
Night time
Mean age 44.5 ± 12.6
years
Severe hypoglycaemic
Study conducted in the
home environment
Monitor-related adverse
event
1
2
reactions
5
*No statistically significant differences in
frequency and duration
32 patients enrolled and
randomised, 27
participated.
There were 1.5 episodes per day of daytime high
subcutaneous glucose (>15 mmol/L), duration
126± 33 minutes, 19.4% of total time.
Mean age 12.5 ± 3.3 years,
range 5-19 years
There were 0.6 episodes per day of night time
high subcutaneous glucose (>15 mmol/L),
duration 177± 83 minutes, 25.5% of total time
16 type-1 diabetic patients
with the CGMS in an open
trial arm and insulin
therapy adjusted
accordingly by team/patient
16 type-1 diabetic patients
with the CGMS in a blinded
study arm with insulin
therapy adjusted based
solely on SMBG
Study conducted in the
home environment
Twenty-six of twenty-seven patients had 0.8
episodes per day of low daytime subcutaneous
glucose (<3.0 mmol/L), duration 58± 29 minutes,
5.5% of total time
All patients had at least one episode per day of
low night time subcutaneous glucose (<3.0
mmol/L), duration 132± 81 minutes, 10.1% of
total time
1 case each of severe hypoglycaemia in each
arm
No statistical significant differences in low
glucose frequency between treatment arms
Sensor functionality curve
3-day 34.9%, 1-2 day 51.7%, none 8.4%
Chico et al
(2003), Spain
Level III-1
evidence
RCT
105 patients enrolled
40 Type-1 diabetic patients
with the CGMS:
Mean age 36.5± 12 years
35 Type-1 diabetic
controls with SMBG:
Asymptomatic
hypoglycaemic events
CGMS
55.7%
(<60 mg/dl) recorded
in subjects
(81 events)
Mean duration 214±288 minutes
Type 1 diabetic patients
Mean age 41± 10 years
Frequency of hypoglycaemic episodes 62.5%
Plus CGMS arm compared
with 30 Type 2 diabetic
patients monitored with
CGMS for asymptomatic
hypoglycemias:
16% during the day, 40% at night, 44% both
Type-2 diabetic patients
Mean age 58± 11 years
Study conducted in the
home environment
Continuous glucose monitoring devices
Frequency of hypoglycaemic episodes 46.6%
42.8% during the day, 42.8% at night, 14.3%
both periods
Adverse events
Misunderstanding instructions
5 patients
Skin lesions
none
Discomfort
8 patients
Sensor malfunction
6 patients
27
Hypoglycaemia and adverse outcomes-The STS™ Continuous Glucose Monitoring
System
One RCT by Garg et al (2006) evaluated the STS™ real-time continuous blood
glucose monitor. Only mild adverse events related to skin complaints were reported
and all resolved within one week and no patients with hypoglycaemic events required
assistance (Table 11).
Table 11 Hypoglycaemic alarm and adverse events and the STS™ system
Study
Population
Outcomes
47 Type-1 and Type-2
diabetic patients wearing
DexCom STS for CGM
(with display data provided
only during 72-hr periods 2
and 3) plus SMBG
There were 21 mild adverse effects from device
use reported in 16 patients.
Features
Garg et al
(2006), USA
Level II
evidence
RCT
44 Type-1 and Type-2
diabetic control patients
wearing DexCom STS for
CGM (with data not
provided during any 72hour period) plus SMBG
These were blister (n=1), bullae around site
(n=1), edema (n=2), erythema (n=17). All
resolved within seven days.
No hypoglycaemic events required assistance in
the display group. Three events (in two
subjects) in control group.
For the 91 subjects mean
age 44± 13 years 75
patients with Type-1
diabetes and 16 with Type2 diabetes
Study conducted in a
clinic/home environment
Summary
Four RCTs reported on safety outcomes for the CGMS® continuous glucose
monitoring system. One small RCT reported significantly more mean hypoglycaemic
episodes detected per subject in the CGMS® group compared to controls and two
other studies reported that asymptomatic hypoglycaemic episodes were commonly
detected in the CGMS® group. Where compared with SMBG, studies showed no
statistically significant differences in the frequency of hypoglycaemic episodes. One
other RCT evaluated the STS™ real-time continuous glucose monitoring system and
minor skin irritations were reported in the CGM device group a greater number of
serious hypoglycaemic events were reported in the control group.
Two RCTs reported on safety outcomes for the GlucoWatch® G2 Biographer. In one
RCT hypoglycaemic events were detected more often with the GlucoWatch® G2
Biographer but this was not able to detect all hypoglycaemic events confirmed by
SMBG and in the other study there was declining usage due to skin reactions which
were common in the GlucoWatch® G2 Biographer group and severe hypoglycaemic
events were at least three times greater in the GlucoWatch® G2 Biographer group of
patients than the SMBG group.
28
Continuous glucose monitoring devices
Potential Cost Impact
Cost Analysis
Findings from the ANZHSN report on the GlucoWatch® G2 Biographer are briefly
repeated here (Australia and New Zealand Horizon Scanning Network, 2005).
Eastman1 et al (2003) (an affiliate of Cygnus international) used a Monte Carlo
simulation model to study the cost-effectiveness of the GlucoWatch® G2 Biographer
based on patients enrolled in the randomised controlled trial described by Chase et al
(2003). This study was conducted in the United States and health costs to the patient
will vary compared to those experienced by Australian and New Zealand patients. It
was assumed that the same frequency of biographer usage would be required for the
lifetime of the patient in order to achieve consistent lowering of HbA1c. The Monte
Carlo model predicted that the use of GlucoWatch® G2 Biographer, if sustained for
life, would delay the onset of the first serious complication of diabetes by 4.1 years.
Treating 18 patients with GlucoWatch® G2 Biographer would prevent one case of
blindness and 1.4 cases of renal failure. However, the validity of the model is
questionable given that there are no long-term morbidity or mortality data reported in
the study by Chase et al. The intervention costs US$91,059 per year of life,
US$61,326 per quality adjust life year (QALY) and US$9,930 per year free of major
complication. If GlucoWatch® G2 Biographer ceased to be effective after 17 years of
age, the cost per QALY would increase to US$103,178 per QALY gained (Eastman et
al 2003).
One other study evaluated the cost-effectiveness of the FreeStyle Navigator™
continuous glucose monitoring system (currently under review with the FDA)
compared with SMBG to predict hypo- and hyper- glycaemic variation in pregnant
women with Type-1 diabetes (Marangos and Papatheofanis, 2005). The study, partly
sponsored by the manufacturer Abbott Laboratories, utilised a Markov model and the
analysis showed that a trained user of the FreeStyle Navigator™ device was more costeffective than SMBG, even though it had a higher overall cost ($US) of treatment for
a 36-month period ($17,305 vs $13,388). CGM use was associated with an improved
quality of life profile compared with SMBG use (53.7 quality-adjusted life months
(QALMs) vs 39.0 QALMs) resulting in a cost-effectiveness ratio of the FreeStyle
Navigator™ of $322/QALM compared with $343 for SMBG. The incremental costeffectiveness ratio (ICER) for the FreeStyle Navigator™ was $267/QALM or
($3,204/QALY) over SMBG. This was also expressed that for an extra 14.7 QALMs
gained the additional cost over SMBG would be $3,917 (Marangos and
Papatheofanis, 2005). The extensive cost of the FreeStyle Navigator™ compared to
SMBG was outweighed by its greater effectiveness. The ICER, and depending on the
willingness-to-pay threshold adopted, the cost/QALY ratio would be extremely
favourable for the FreeStyle Navigator™ over SMBG. The limitations with this study
are that the study was conducted in the United States so health costs will be different
to those experienced by Australian and New Zealand patients. The patient data used
in the model was preliminary and sourced from on-going trials and the model itself
had limitations in terms of representing patients with multiple diabetes complications,
1
Dr Richard Eastman is affiliated with Cygnus Incorporated
Continuous glucose monitoring devices
29
and assuming complete patient compliance and effectiveness of treatments in
returning patients to a stable euglycaemic state.
Simple costings
Medtronic Australasia Pty. Ltd markets the CGMS® and its successor the CGMS®
System Gold™ continuous glucose monitoring system. In Australia this system costs
approximately A$5,800 and a box of 4 or 10 sensors costs A$300 – A$700. A new
AutoSensor is required for every 12 hours of monitoring. In addition the inserter
required by the patient for device attachment costs A$120 (Medtronic Australasia,
personal communication). These are ineligible for a subsidy from the National
Diabetic Supply Scheme (NDSS). In New Zealand CGMS® models are distributed
through Medica Pacifica with a cost of approximately NZ$8,000 for the system and
NZ$78 per sensor. No other continuous glucose monitoring systems are currently
available in Australia or New Zealand. By contrast the GlucoWatch® G2 Biographer
if purchased in either the United States or the United Kingdom would cost
approximately A$900 and a packet of 16 AutoSensors (one use only) is A$130, or
A$8 each. A new AutoSensor is required for every 13 hours of monitoring (McGahan
2002).
Continuous blood monitoring systems are an adjunct to and do not replace standard
finger-prick SMBG. Although continuous monitoring devices may reduce the number
of finger-prick blood glucose tests patients will still be required to purchase a blood
glucose monitor and strips. A blood glucose monitor such as the Roche Diagnostic
Accu-Chek Advantage 3 currently costs A$70 and has a lifetime guarantee (Roche
Diagnostics Australia Pty Limited). Newly diagnosed diabetic patients are issued
with a National Diabetic Supply Scheme (NDSS) card. The NDSS is an Australian
Government registration scheme, which provides a subsidy for blood glucose testing
strips, free insulin syringes and free needles for insulin delivery pens. The NDSS does
not provide a subsidy for blood glucose meters, lancets or lancet devices. By quoting
their unique NDSS number, patients may order testing strips from their local diabetic
association. Testing strips and lancets currently cost approximately A$13 (packet 100)
and A$16 (packet of 200), respectively and would cost a total of approximately 64
cents per day if patients tested three times daily (personal communication, Diabetes
South Australia).
Ethical Considerations
Informed Consent
In the clinical studies reviewed patients (and the parents of patients where required)
provided informed written consent to participate and study protocols underwent ethics
approval. Subjects were informed of the device specifications, limitations and
training in their use was provided. In clinical practice patients should be provided
with this information for each CGM device. CGM is not a substitute for conventional
point-in-time blood glucose testing but may be offered as an adjunct to assist with
self-monitoring and glycaemic control and to reduce the number of daily conventional
tests. These devices assist in achieving near normoglycaemia, while minimizing the
risk of unexpected hypoglycaemia.
30
Continuous glucose monitoring devices
Access Issues
Commercial CGM devices are being produced by a growing number of companies
and approved for clinical use, particularly in the U.S. Currently these devices are
available on a prescription basis and patient use is closely monitored. There is only
limited availability of these devices in Australia and New Zealand and their use is
strictly controlled.
Companies will need to create more affordable and viable CGM products if there is to
be greater acceptance and utilisation of CGM among clinicians and patients. Critical
factors that may determine this include the development of systems with real-time
capability and being comfortable to wear for patients. The development of
performance standards that are diagnosis specific, combine accuracy for point and
trend prediction, provide definitions for varying magnitudes of glycaemia, and device
specific performance standards will provide target specifications that will ensure
better CGM devices (Klonoff, 2005b). Other factors that will promote the
development of CGM technologies in clinical practice are the provision of more
outcomes data related to improvements in HbA1c, reductions in the frequency and
severity of hypoglycaemic episodes, clinical guidelines for the use of CGM, and
algorithms and performance assessment for real-time readings compared with
retrospective analysis (Klonoff, 2005b).
The potential uptake of CGM in clinical practice is huge given the clinical need and
burden of disease associated with diabetes mellitus worldwide and the benefits for
patients requiring blood glucose monitoring, especially for those with unstable
glycaemic control and for paediatric patients. At present the potential utilisation of
CGM will remain as an adjunctive procedure and will not replace existing SMBG
methods until limitations in non-invasive and minimally invasive CGM technologies
are resolved.
Training and Accreditation
Training
Patients and the parents/care givers of patients require training in the use of
continuous blood glucose monitoring systems. Within Australia and New Zealand
these can only be used in controlled physician-supervised settings. Apart from
product inserts containing instructions for using these systems company
representatives are the main source of training. Physicians and diabetes nurse
educators are the primary recipients of company training, who in turn, would train
diabetes patients and their parents/care givers on device use. CGM systems are an
adjunct to and do not replace standard finger-prick and home glucose meter testing.
Clinical Guidelines
Recent Australian clinical practice guidelines for children and adolescents with Type1 diabetes have been published (Department of Health and Aging, 2005a). These
guidelines address the diagnosis and clinical management of Type-1 diabetes in
children of all ages, including adolescents up to the point of transition to adult care.
Related to glycaemic control the guidelines recommend (1) that diabetes control
Continuous glucose monitoring devices
31
should be optimised as much as possible as improved glycaemic control reduces the
risk of development and progression of microvascular and macrovascular
complications in adolescents and adults. (2) Frequent daily blood glucose monitoring
as part of a package of care has been shown to be associated with improved glycaemic
control. (3) The frequency of blood glucose monitoring should be adapted to the
insulin regimen, the age of the child and the stability of the diabetes. (4) HbA1c is the
only measure of glycaemic control that has been shown to be associated with longterm complications of diabetes and best reflects glycaemic levels over the preceding
2-3 months. (5) The American Diabetes Association recommends measuring the
HbA1c at least twice per year in patients who are meeting treatment goals, and more
frequently (quarterly) in those whose treatment has changed or who are not meeting
glycaemic goals.
In the guidelines minimally invasive continuous glucose monitoring systems are
recognised as a useful tool for detecting asymptomatic hypoglycaemia and for
providing detailed information on blood glucose trends during stabilisation of diabetes
or during initiation and monitoring of insulin pump therapy. There is only limited
introduction of the minimally invasive systems (e.g. the CGMS® system) within
Australia and New Zealand currently and these are only available for short-term use in
individual patients attending larger hospitals and clinics. Non-invasive devices (e.g.
GlucoWatch® G2 Biographer) have not been released in Australia and New Zealand.
Recommended frequencies for SMBG to optimise or advance therapy range from a
recommended 1 to 4 times daily depending upon the type of therapy, the degree of
glycaemic control, risk of hypoglycaemia, the need for short term treatment
adjustments and other special situations such as pregnancy (Bergenstal and Gavin,
2005).
The timing of SMBG should be at various times during the day, including
preprandially and 1 to 2 hours postprandially. The value of SMBG for Type-2
diabetes patients not treated with insulin has been the subject of some debate in the
literature but recent evidence supports its use in these patients (Bergenstal and Gavin,
2005).
There are also available recently published Australian evidence based guidelines for
the primary prevention, case detection and diagnosis of Type 2 diabetes (Department
of Health and Aging, 2005b).
Sources of Further Information
Two Horizon Scanning reports on CGM devices have been produced by the Spanish
Agency for Health Technology Assessment in Andalusia (AETSA).
One assessment considered RCTs and CCTs on the Cygnus GlucoWatch® G2
Biographer (AETSA, 2005a). The key findings were that most studies were of low to
poor quality, there was good correlation with standard monitoring, accuracy decreased
with increasing glucose levels, and sensitivity for hypoglycaemia increased with preset detection levels but so did the false positive rate. There were contradictory results
in terms of blood glucose control (measured by HbA1c levels), there was no
32
Continuous glucose monitoring devices
improvement in quality of life compared with SMBG and skin irritations were the
main adverse event and resolved after discontinuation. One economic assessment
study was found.
The other horizon scanning assessment considered RCTs and CCTs on the CGMS®
and Glucoday® minimally invasive CGM systems (AETSA, 2005b). The key findings
were that most studies were of low to poor quality, there was good correlation with
standard monitoring, accuracy was satisfactory in the euglycaemic range but tended to
overestimate the frequency and duration of hypoglycaemic events, contradictory
results were seen in the improvement of blood glucose control (measured by HbA1c
levels), there was no improvement in quality of life compared with SMBG and minor
skin irritations were the main adverse event. No economic assessment studies or
long-term data were found.
A number of non-randomised controlled studies were also identified on the
GlucoWatch® G2 Biographer and the CGMS® Continuous Glucose Monitoring
System. These were the first clinically approved and commercially available devices
and have been studied more extensively in the medical literature than other devices.
Several RCTs investigating various continuous glucose monitoring devices are listed
here. These trials are on-going or have recently been completed.
An ongoing U.K. based four-arm multi-centre RCT (Minimally Invasive Technology
Role and Evaluation- MITRE trial) with a sample size of 600 subjects aged 18+ years
looking at the Cygnus GlucoWatch® G2 Biographer, Medtronic MiniMed CGMS®,
attention control with a frequency of nurse feedback sessions the same as the CGM
device groups, and conventional monitoring (one visit every six months) in the
management of insulin treated diabetes mellitus (National Research Register
Document – N0484119008).
An ongoing U.K. based multi-centre RCT evaluating the Medtronic MiniMed
CGMS® for continuous monitoring in 120 subjects with pregnancies complicated by
pre-existing diabetes. Subjects were allocated to either standard antenatal care or
CGMS® in addition to standard care (National Research Register Document –
N0254145814). One other U.K. based single-centre RCT similarly examining the
CGMS® in 30 subjects with pregnancies complicated by pre-existing diabetes is ongoing (National Research Register Document – N0547148012).
The Guardcontrol Trial is an international multi-centre RCT to assess whether insulin
dependent Type-1 diabetes patients with poor glycaemic control can be improved by
the utilisation of the Guardian®RT Telemetered Glucose Monitoring System
compared with finger-stick based self testing (SMBG). Expected enrolment was 162
subjects and this trial was recently completed (ClinicalTrials.gov identifier
NCT00111228).
Impact Summary
Due to the importance of optimizing glycaemic control in diabetes management and
limitations with conventional SMBG methods there has been considerable investment
and development in CGM technologies with more than 100 companies currently
Continuous glucose monitoring devices
33
involved in the research. At present there are no large-scale clinical studies
supporting the use of non-invasive CGM technologies that use the interaction of
electromagnetic radiation with tissue in glucose detection. Therefore, minimally
invasive technologies that sample and monitor glucose concentrations in the ISF
across the skin offer the greatest potential for practical CGM in clinical practice.
Seven CGMs have been approved by the FDA for clinical use in the U.S. or carry the
CE marking for clinical use in Europe. The GlucoWatch® G2 Biographer and the
original CGMS® Continuous Glucose Monitoring System and its second-generation
replacement the CGMS® System Gold™ were the first clinically approved and
commercially available devices and have been studied more extensively in the
medical literature than other devices. Only the CGMS® Continuous Monitoring
System and its successor the CGMS® System Gold™ are currently available in
Australia and New Zealand and access is strictly controlled through physiciansupervised use in the management of specialised diabetes cases. In Australia the
devices are currently not approved by the TGA.
The potential uptake for CGM is huge given the clinical need and burden of disease
associated with diabetes mellitus worldwide and because testing glucose levels is
essential for children, adolescents and adults who suffer from Type-1 and Type-2
diabetes requiring treatment with insulin and for people taking oral hypoglycaemic
agents. CGM is an adjunct to standard finger-prick SMBG and current CGM
technology will not replace this. Currently there is a need to develop more affordable
and viable CGM devices with sound performance standards and produce more
evidence of efficacy and safety in clinical trials if there is to be greater acceptance and
utilisation of CGM among clinicians and patients.
Conclusions
The importance of achieving optimal glycaemic control in diabetes management and
limitations with conventional SMBG methods led to the development of continuous
glucose monitoring (CGM) technologies with more than 100 companies currently
involved in the research. At present there are no large-scale clinical studies
supporting the efficacy, portability, and affordability of non-invasive technologies.
Therefore, minimally invasive technologies are offering the greatest potential for
practical continuous glucose monitoring devices for clinical use. At this stage only
non-invasive impedance spectroscopy (the application of electromagnetic radiation
through the skin to the blood vessels) and minimally invasive technologies (interstitial
fluid (ISF) measurement in situ or extraction through the skin) are approved for
clinical use.
CGM systems consist of a small monitor that reads and displays glucose values in
real-time or retrospectively, a glucose sensor which is implanted subcutaneously in
the abdomen (or externally to the wrist, arm or forearm), and a transmitter to relay
information about glucose concentrations between the sensor and monitor. Typically,
each device undergoes a warm-up period of 1-2 hours, a device specific calibration
process of between 1 and 4 times per day and each device’s sensor provides a blood
glucose reading every 1-10 minutes for up to 72 hours and up to 3-months for newer
non-invasive technology. The glucose level information is available to the patient and
34
Continuous glucose monitoring devices
clinician either in real time (a feature of more recently developed models) or
retrospectively and many models have alarms that trigger if glucose levels fall outside
of preset euglycaemic ranges.
Seven CGMs have been approved by the U.S. Food and Drug Administration (FDA)
for clinical use in the U.S. or carry the CE marking for clinical use in Europe. The
Cygnus GlucoWatch® G2 Biographer and the original Medtronic MiniMed CGMS®
Continuous Glucose Monitoring System and its second-generation replacement the
CGMS® System Gold™ were the first clinically approved and commercially available
devices. These devices have been studied more extensively in the literature than other
devices. The RCT studies in this report mainly utilised the original CGMS®, the
GlucoWatch® G2 Biographer and in one RCT, the Guardian® RT Continuous Glucose
Monitoring System and in another the STS™ Continuous Glucose Monitoring System.
Only the CGMS® Continuous Monitoring System and its successor the CGMS®
System Gold™ are currently available for limited use in Australia and New Zealand
(currently not approved by the TGA in Australia) but are only available in approved
institutions for controlled physician-supervised use in the management of specialised
diabetes cases.
The accuracy of the original CGMS® and its modified successor the CGMS® System
Gold™ and the GlucoWatch® G2 Biographer have been rigorously assessed in clinical
studies. These studies found that both sensors were less accurate during
hypoglycaemia and that gender, ethnicity, Body Mass Index (BMI), or age (3-18
years) had no effect on the function of either sensor. The CGMS® sensors were
equally accurate on each of the three days of wear while the GlucoWatch® G2
Biographer was less accurate in the last 5-hours of wear. In comparing day and nighttime accuracy the GlucoWatch® G2 Biographer showed no differences in accuracy
whilst the CGMS® was less accurate with lower readings at night.
Diagnostic accuracy outcomes for CGM devices were reported in four RCT studies
(level 3b evidence in the assessment of diagnostic accuracy hierarchy). Three RCTs
reported diagnostic accuracy in the Medtronic MiniMed Guardian® and the CGMS®
glucose sensors. In these studies the CGMS® sensors were less accurate than home
sensors (SMBG) and less accurate during periods of hypoglycaemia. One RCT
reported on the accuracy of real-time sensor values for the Dexcom STS™ Continuous
Glucose Monitoring System compared with SMBG values. Overall 95.4% of paired
glucose values were within Clarke error grid A and B zones with a correlation
coefficient of 0.88. There were mean and median absolute differences but sensor
values were within pre-specified accuracy limits with SMBG glucose values and no
systematic bias detected.
Effectiveness was evaluated in four RCTs (two level II and two with poorer quality
study designs with level III-1 evidence) with glycaemic control outcomes using the
CGMS® continuous glucose monitoring system compared with SMBG. In the two
studies with paediatric patients therapy adjustments were made for patients on the
basis of CGMS® and resulted in improvements in HbA1c levels. In the other two
studies with adult patients there were significant improvements in HbA1c levels in
both the CGMS® and control groups. Two of the RCTs did not provide betweengroup comparisons. Where these were compared in one study both groups had similar
improvement while in the other the CGMS® group had greater improvement in
HbA1c levels.
Continuous glucose monitoring devices
35
Effectiveness was also evaluated in two RCTs (level II evidence) with glycaemic
control outcomes for patients with Type-1 diabetes using the GlucoWatch® G2
Biographer compared with SMBG. One study found a significant improvement in
HbA1c levels in the CGM monitoring group at three months but not at six and nine
months follow-up and the other study found no significant decreases or differences in
either group in HbA1c at six months.
One RCT (level II evidence) evaluated the effectiveness of the Guardian® CGMS and
found the glucose sensor alert group had significantly reduced median hypoglycaemia
excursion periods compared to the control group without sensor alert. One other RCT
(level II evidence) evaluated the STS™ continuous glucose monitoring system and
found that real-time sensor monitoring assisted in significant reductions in periods of
hypoglycaemia and hyperglycaemia.
In the three RCTs considered there was a lack of clear evidence indicating significant
improvements in quality of life for patients using CGM devices compared to SMBG
as assessed by various psychometric measures.
Four RCTs (two level II and two with poorer quality study designs with level III-1
evidence) reported on safety outcomes for the CGMS® continuous glucose monitoring
system compared to a control group using SMBG. In one small RCT there were
significantly more mean hypoglycaemic episodes detected per subject in the CGMS®
group compared to controls and two studies reported that asymptomatic
hypoglycaemic episodes were commonly detected in the CGMS® group. Where
compared, two studies showed no statistically significant differences in the frequency
of hypoglycaemic episodes between the two study arms. Asymptomatic
hypoglycaemic episodes were reported in the CGMS® groups. One other RCT
evaluated the STS™ real-time continuous glucose monitoring system and minor skin
irritations were reported in the CGM device group. There were a greater number of
serious hypoglycaemic events in the control group.
Two RCTs (level II evidence) reported on safety outcomes for the GlucoWatch® G2
Biographer compared to a control group using SMBG. In one RCT hypoglycaemic
events were detected more often with the GlucoWatch® G2 Biographer but this was
not able to detect all hypoglycaemic events confirmed by SMBG and in the other
study there was declining usage due to skin reactions which were common in the
GlucoWatch® G2 Biographer group and severe hypoglycaemic events were at least
three times greater in the GlucoWatch® G2 Biographer group of patients than the
SMBG group.
The quality of the evidence in these RCTs was limited by a lack of between-group
comparison with patients in the SMBG group and the very select (mainly children and
adolescents) and small patient samples assessed in the studies.
There were two cost-effectiveness studies identified, one evaluated the GlucoWatch®
G2 Biographer, and the other evaluated the FreeStyle™ Navigator system. It was
estimated from a Monte Carlo simulation that life time use of the GlucoWatch® G2
Biographer would delay the onset of serious diabetes complications by 4.1 years and
treating 18 patients would prevent one case of blindness and 1.4 cases of renal failure.
The intervention cost was US$61,326 per QALY and US$9,930 per year free of major
complication. However the major limitation was that no long-term morbidity and
mortality data were included in the model. The other study showed that the high cost
of the FreeStyle™ Navigator compared to SMBG was outweighed by its greater
effectiveness. The incremental cost-effectiveness ratio (ICER), and depending on the
36
Continuous glucose monitoring devices
willingness-to-pay threshold adopted, the cost/QALY ratio may be favourable for the
FreeStyle Navigator™ over SMBG. However, there were limitations with this study
as the model data were preliminary and broad assumptions were made about patient
conditions and treatment effectiveness.
Testing blood glucose levels is essential for children, adolescents and adults who
suffer from Type-1 and Type-2 diabetes requiring treatment with insulin and for
people taking oral hypoglycaemic agents. The potential uptake of CGM in clinical
practice is huge given the clinical need and burden of disease associated with diabetes
mellitus worldwide and the benefits for patients requiring blood glucose monitoring,
especially for those with unstable glycaemic control and for paediatric patients.
Current CGM systems are an adjunct to SMBG and do not replace standard fingerprick testing in SMBG. Although continuous monitoring devices may reduce the
number of finger-prick blood glucose tests patients will still be required to purchase a
blood glucose monitor and strips. There is a need to develop more affordable and
viable CGM products with real-time capability and maximum comfort if there is to be
greater acceptance and utilisation of CGM among clinicians and patients. Other
factors that will promote the development of CGM technologies in clinical practice
are the development of performance standards and the provision of more outcomes
data related to improvements in HbA1c, reductions in the frequency and severity of
hypoglycaemic episodes, clinical guidelines for the use of CGM, and algorithms and
performance assessment for real-time readings compared with retrospective analysis.
HPACT Advisory:
There is significant potential for the uptake of continuous glucose monitoring (CGM)
devices given the worldwide clinical need and burden of disease associated with
diabetes mellitus. There is a need to develop more affordable and viable CGM devices
with sound performance standards and to show more beneficial clinical effectiveness
and safety outcomes if there is to be greater acceptance and utilisation of CGM
devices. Evidence from RCTs, though somewhat contradictory and limited by small
and select patient groups, indicates some effectiveness in glycaemic control and
increased safety due to greater awareness of glycaemic variation but these devices are
less accurate, particularly during hypoglycaemic episodes and can cause minor skin
reactions, and do not improve diabetes related quality of life, compared with SMBG.
CGM is useful as an adjunct to conventional (standard blood glucose self-monitoring)
SMBG in selected patients with difficulties in maintaining glycaemic control.
However, at this stage, CGM will not replace conventional SMBG in the majority of
patients.
Continuous glucose monitoring devices
37
Appendix: Levels of Evidence
There were thirteen published studies included for assessment in this report. All studies
were graded according to the dimensions of evidence defined by the National Health
and Medical Research Council (NHMRC, 1999) (Table 12) and/or levels of evidence
for assessing diagnostic accuracy (Phillips et al 2001) (Table 13).
Four RCTs reported diagnostic accuracy outcomes (level 3b evidence). There were two
level II evidence grade RCTs which reported safety, effectiveness outcomes related to
glycaemic control for the GlucoWatch® G2 Biographer and one of these studies plus
one other RCT reported on quality of life outcomes for paediatric patients. Four RCTs
(two level II and two level III-1 evidence) reported effectiveness and safety outcomes
related to glycaemic control in paediatric and adult patients utilising the CGMS®
continuous glucose monitoring system. One of these RCTs reported on quality of life
outcomes for the CGMS®. One further RCT reported effectiveness outcomes for the
Guardian® CGMS. One RCT reported on effectiveness and safety outcomes in adult
patients utilising the real-time STS™ continuous monitoring system. Three studies,
Fiallo-Scharer et al (2005), Chase et al (2005) and The Diabetes Research in Children
Network (DirecNet) Study group (2006) were part of the same trial.
Two cost-effectiveness studies were also included with one RCT based analysis for the
GlucoWatch® G2 Biographer and the other for FreeStyle Navigator™ continuous
glucose monitoring device.
Of the thirteen studies included, two studies were sponsored by and at least two authors
were employees of Medtronic MiniMed manufacturing the CGMS device. In three
other studies one author was an employee of Cygnus Inc and in one study an employee
of Medtronic MiniMed. In four studies monitoring devices were provided free by the
manufacturer. All studies, except one that was conducted in Sweden, were conducted
within the USA.
Table 12 Designations of levels of evidence
Level of
Study design
evidence
I
Evidence obtained from a systematic review of all relevant randomised controlled trials
II
Evidence obtained from at least one properly-designed randomised controlled trial
III-1
Evidence obtained from well-designed pseudorandomised controlled trials (alternate
allocation or some other method)
III-2
Evidence obtained from comparative studies (including systematic reviews of such
studies) with concurrent controls and allocation not randomised, cohort studies, casecontrol studies, or interrupted time series with a control group
III-3
Evidence obtained from comparative studies with historical control, two or more single
arm studies, or interrupted time series without a parallel control group
IV
Evidence obtained from case series, either post-test or pre-test/post-test
Modified from: National Health and Medical Research Council (1999). A guide to the development, implementation and
evaluation of clinical practice guidelines, Commonwealth of Australia, Canberra, ACT.
38
Continuous glucose monitoring devices
Table 13 Levels of evidence for assessing diagnostic accuracya
Level of
evidence
1a
Study design
1b
Validating** cohort study with good† reference standards; or CDR tested within one
clinical centre
1c
Absolute SpPins and SnNouts
2a
SR (with homogeneity*) of Level ≥2 diagnostic studies
2b
Exploratory** cohort study with good† reference standards; CDR after derivation, or
validated only on split-sample§ or databases
2c
n/a
3a
SR (with homogeneity*) of 3b and better studies
3b
Non-consecutive study; or without consistently applied reference standards
4
Case-control study, poor or non-independent reference standard
5
Expert opinion without explicit critical appraisal, or based on physiology, bench
research or “first principles”
a
SR (with homogeneity*) of Level 1 diagnostic studies; CDR with 1b studies from
different clinical centres
††
(Phillips et al 2001). SR = systematic review; CDR = clinical decision rule - these are algorithms or scoring systems
which lead to a prognostic estimation or a diagnostic category; RCT = randomised controlled trial; n/a = not
applicable.
* Homogeneity means a systematic review that is free of worrisome variations (heterogeneity) in the directions and
degrees of results between individual studies. Not all systematic reviews with statistically significant heterogeneity
need be worrisome, and not all worrisome heterogeneity need be statistically significant. Studies displaying
worrisome heterogeneity should be tagged with a “-“ at the end of their designated level. ** Validating studies test the
quality of a specific diagnostic test, based on prior evidence. An exploratory study collects information and trawls the
†
data (e.g. using a regression analysis) to find which factors are 'significant'. Good reference standards are
independent of the test, and applied blindly or objectively to all patients. Poor reference standards are haphazardly
applied, but still independent of the test. Use of a non-independent reference standard (where the 'test' is included in
††
the 'reference', or where the 'testing' affects the 'reference') implies a level 4 study. An “Absolute SpPin” is a
diagnostic finding whose Specificity is so high that a Positive result rules-in the diagnosis. An “Absolute SnNout” is a
§
diagnostic finding whose Sensitivity is so high that a Negative result rules-out the diagnosis. Split-sample validation
is achieved by collecting all the information in a single tranche, then artificially dividing this into "derivation" and
"validation" samples.
Search Strategy
The medical literature (Table 14) was searched utilising the search terms outlined
(Table 15) to identify relevant studies and reviews, until 15th March 2006. In
addition, major international health technology assessment databases were searched.
Table 14 Literature sources utilised in assessment
Source
Electronic databases
Cinahl
Cochrane Library – including, Cochrane Database of Systematic
Reviews, Database of Abstracts of Reviews of Effects, the Cochrane
Central Register of Controlled Trials (CENTRAL)
Current Contents
Embase
Pre-Medline and Medline
Medline (via PubMed last 60 days)
International Pharmaceutical Abstracts
Web of Science
The Health Technology Assessment Database, the NHS Economic
Evaluation Databases
Location
Ovid
Ovid
ISI
Ovid
Ovid
National Library of Medicine
Ovid
ISI
http://www.york.ac.uk/inst/crd/crddatabases.htm
Internet
Continuous glucose monitoring devices
39
American Diabetes Association
US Food & Drug Administration (FDA) Center for Devices &
Radiological Health
Australian Institute of Health and Welfare
Australian NH&MRC
Statistics New Zealand
New Zealand Ministry of Health
Canadian Coordinating Office for Heath Technology Assessment
Emerging Technology List
Blue Cross Blue Shield Technology Evaluation Center
Agencia de Evaluacion de Technologias Sanitarias de Andalusia
Google Search Engine and Google Scholar Search
www.diabetes.org
http://www.fda.gov/cdrh
http://www.aihw.gov.au/
http://www.nhmrc.gov.au/
http://www.stats.govt.nz/
http://www.moh.govt.nz/
http://www.ccohta.ca
http://www.bcbs.com/tec/tecassessments.html
http://www.juntadeandulucia.es/salud/orgdep/aetsa/
http://scholar.google.com
Table 15 Search terms utilised
Search terms
MeSH & EMBASE headings
N/A
Text words
(Continuous adj3 glucose), ((continuous adj3 monitor$) and glucose), minimed, glucowatch, gluco watch,
pendra, freestyle navigator, glucoday, guardian RT, (guardian and glucose), STS system
AND
Random$, control$, trial$, (systematic$ adj3 review$)
Limits
No limits by date or language
Limitations of the Assessment
Methodological issues and the relevance or currency of information provided over
time are paramount in any assessment carried out in the early life of a technology.
Horizon Scanning forms an integral component of Health Technology Assessment.
However, it is a specialised and quite distinct activity conducted for an entirely
different purpose compared to a comprehensive systematic review. The rapid
evolution of technological advances can in some cases overtake the speed at which
trials or other reviews are conducted. In many cases, by the time a study or review has
been completed, the technology may have evolved to a higher level leaving the
technology under investigation obsolete and replaced.
An Horizon Scanning Report maintains a predictive or speculative focus, often based
on low level evidence, and is aimed at informing policy and decision makers. It is not
a definitive assessment of the safety, effectiveness, ethical considerations and cost
effectiveness of a technology.
In the context of a rapidly evolving technology, an Horizon Scanning Report is a
‘state of play’ assessment that presents a trade-off between the value of early,
uncertain information, versus the value of certain, but late information that may be of
limited relevance to policy and decision makers.
This report provides an assessment of the current state of development of continuous
glucose monitoring devices, their present and potential use in the Australian and New
Zealand public health systems, and future implications for the use of this technology.
40
Continuous glucose monitoring devices
References
Agency for Health Technology Assessment in Andalusia (AETSA) (2005a).
Continuous blood glucose monitoring system “GlucoWatch®”. Report 2/2005.
Seville, AETSA.
Agency for Health Technology Assessment in Andalusia (AETSA) (2005b).
Minimally invasive systems for the continuous glucose monitoring system. Report
3/2005. Seville, AETSA.
American Diabetes Association (2004). Standards of medical care in diabetes.
Diabetes Care, 27 (Suppl 1):S15-S35.
Aussedat, B., Dupire-Angel, M., Gifford, R., Klein, J., Wilson, G., & Reach, G.
(2000). Interstitial glucose concentration and glycaemia: implications for continuous
sub-cutaneous glucose monitoring. American Journal of Physiology Endocrinology
and Metabolism, 278, E716-E728.
Australia and New Zealand Horizon Scanning Network. (2005). GlucoWatch® G2
Biographer for the non-invasive monitoring of glucose levels. Adelaide Health
Technology Assessment.
Australian Institute of Health and Welfare (2002). Diabetes: Australian facts 2002.
AIHW Cat. No. CVD 20. Canberra: AIHW
Australian Institute of Health and Welfare (2003). National Diabetes Register:
statistical profile, December 2001. AIHW Cat. No. CVD 24. Canberra: AIHW.
Australian Institute of Health and Welfare (2004). Australia’s Health 2004.
AIHW Cat. No. AUS 44. Canberra: AIHW.
Australian Institute of Health and Welfare: Dixon, T. (2005a). Diabetes related
deaths in Australia, 2001-2003. Bulletin No. 32. AIHW Cat No. AUS 69. Canberra:
AIHW.
Australian Institute of Health and Welfare: Dixon, T. (2005b). Costs of diabetes in
Australia, 2000-01. Bulletin No. 26. AIHW Cat. No. AUS 59. Canberra: AIHW.
Babu, A. R., Herdegen, J., Fogelfeld, L., Shott, S., & Mazzone, T. (2005). Type 2
diabetes, glycemic control, and continuous positive airway pressure in obstructive
sleep apnea. Archives of Internal Medicine, 165, 447-452.
Beardsall, K., Ogilvy-Stuart, A. L., Ahluwalia, J., Thompson, M., & Dunger, D. B.
(2005). The continuous glucose monitoring sensor in neonatal intensive care. Archives
of Disease in Childhood Fetal & Neonatal Edition, 90, F307-310.
Bergenstal, R., and Gavin, J. (2005). The role of self-monitoring of blood glucose in
the care of people with diabetes: report of a global consensus conference. Global
Continuous glucose monitoring devices
41
Consensus Conference on Glucose Monitoring Panel. The American Journal of
Medicine, 118, 1S-6S.
Bode, B., Gross, K., Rikalo, N., Schwartz, S., Wahl, T., Page, C., Gross, T., et al.
(2004). Alarms based on real-time sensor glucose values alert patients to hypo- and
hyperglycemia: the guardian continuous monitoring system. Diabetes Technology &
Therapeutics, 6, 105-113.
Braunwald, E., Fauci, A. et al (Eds.) (2001). Harrison's Principles of Internal
Medicine, McGraw-Hill, New York.
Buckingham, B., Block, J., & Wilson, D. M. (2005). Continuous glucose monitoring.
Current Opinion in Endocrinology & Diabetes, 12, 273-279.
Buhling, K. J., Kurzidim, B., Wolf, C., Wohlfarth, K., Mahmoudi, M., Wascher, C.,
Siebert, G., et al. (2004). Introductory experience with the continuous glucose
monitoring system (CGMS; Medtronic Minimed) in detecting hyperglycemia by
comparing the self-monitoring of blood glucose (SMBG) in non-pregnant women and
in pregnant women with impaired glucose tolerance and gestational diabetes.
Experimental & Clinical Endocrinology & Diabetes, 112, 556-560.
Bui, H., Perlman, K., & Daneman, D. (2005). Self-monitoring of blood glucose in
children and teens with diabetes. Paediatric Diabetes, 6, 50-62.
Cameron, F. J., & Ambler, G. R. (2004). Does continuous glucose monitoring have
clinical utility in contemporary management of diabetes? Journal of Paediatrics &
Child Health, 40, 79-84.
Chase, H. P., Kim, L. M., Owen, S. L., MacKenzie, T. A., Klingensmith, G. J.,
Murtfeldt, R., & Garg, S. K. (2001). Continuous subcutaneous glucose monitoring in
children with type 1 diabetes. Pediatrics, 107, 222-226.
Chase, H. P., Roberts, M. D. et al (2003). Use of the GlucoWatch biographer in
children with type 1 diabetes, Pediatrics, 111 (4 Pt 1), 790-794.
Chase, H. P., Beck, R., Tamborlane, W., Buckingham, B., Maurus, N., Tsalikian, E.,
et al. (2005). A randomised multicenter trial comparing the GlucoWatch Biographer
with standard glucose monitoring in children with Type 1 diabetes. Diabetes Care,
28, 1101-1106.
Chico, A., Vidal-Rios, P., Subira, M., & Novials, A. (2003). The continuous glucose
monitoring system is useful for detecting unrecognized hypoglycemias in patients
with type 1 and type 2 diabetes but is not better than frequent capillary glucose
measurements for improving metabolic control. Diabetes Care, 26, 1153-1157.
Choleau, C., Klein, J., Reach, G., Aussedat, B., Demaria-Pesce, V., Wilson, G.,
Gifford, R., & Ward, W. (2002). Calibration of a subcutaneous amperometric glucose
sensor. Part 1. Effect of measurement uncertainties on the determination of sensor
sensitivity and background current. Biosensors and Bioelectronics, 17, 641-646.
42
Continuous glucose monitoring devices
Conrad, S. C., Mastrototaro, J. J., & Gitelman, S. E. (2004). The use of a continuous
glucose monitoring system in hypoglycemic disorders. Journal of Pediatric
Endocrinology, 17, 281-288.
Czupryniak, L., Loba, J., Pawlowski, M., Nowak, D., & Bialasiewicz, P. (2005).
Treatment with continuous positive airway pressure may affect blood glucose levels
in nondiabetic patients with obstructive sleep apnea syndrome. Sleep, 28, 601-603.
Department of Health and Aging (2005a). Clinical Practice Guidelines: Type 1
Diabetes in Children and Adolescents. National Health and Medical Research
Council. Available from http://www.nhmrc.gov.au/publications/_files/cp102.pdf
Accessed on 7.04.06.
Department of Health and Aging (2005b). Evidence based guidelines for Type-2
diabetes: Primary prevention, case detection, diagnosis. National Health and Medical
Research Council. Available from
http://www.nhmrc.gov.au/publications/_files/cp86.pdf Accessed on 7.04.06.
Eastman, R. C., Leptien, A. D. & Chase, H. P. (2003). Cost-effectiveness of use of the
GlucoWatch™ Biographer in children and adolescents with type 1 diabetes: A
preliminary analysis based on a randomized controlled trial, Pediatric Diabetes, 4 (2),
82-86.
Federal Dug Administration (2002). Diabetes information. Available from:
http://www.fda.gov/diabetes/glucose.html Accessed on 10.04.06.
Fiallo-Scharer, R., & Diabetes Research in Children Network Study, G. (2005). Eightpoint glucose testing versus the continuous glucose monitoring system in evaluation
of glycemic control in type 1 diabetes. Journal of Clinical Endocrinology &
Metabolism, 90, 3387-3391.
Garg, S., Zisser, H., Schwartz, S., Bailey, T., Kaplan, R., Ellis, S., & Jovanovic, L.
(2006). Improvement in glycemic excursions with a transcutaneous, real-time
continuous glucose sensor: a randomized controlled trial. Diabetes Care, 29, 44-50.
Hershkovitz, E., Rachmel, A., Ben-Zaken, H., & Phillip, M. (2001). Continuous
glucose monitoring in children with glycogen storage disease type I. Journal of
Inherited Metabolic Disease, 24, 863-869.
Hirsch, I. B., Bode, B. W., Garg, S., Lane, W. S., Insulin Aspart, C. M. D. I. C., & et
al. (2005). Continuous subcutaneous insulin infusion (CSII)of insulin aspart versus
multiple daily injection of insulin aspart/insulin glargine in type 1 diabetic patients
previously treated with CSII. Diabetes Care, 28, 533-538.
Hovorka, R. (2006). Continuous glucose monitoring and closed-loop systems.
Diabetic Medicine, 23, 1-12.
Javid, P. J., Halwick, D. R., Betit, P., Thompson, J. E., Long, K., Zhang, Y., Jaksic,
T., et al. (2005). The first use of live continuous glucose monitoring in patients on
extracorporeal life support. Diabetes Technology & Therapeutics, 7, 431-439.
Continuous glucose monitoring devices
43
Kerssen, A., de Valk, H. W., & Visser, G. H. (2004). The Continuous Glucose
Monitoring System during pregnancy of women with type 1 diabetes mellitus:
accuracy assessment. Diabetes Technology & Therapeutics, 6, 645-651.
Kerssen, A., De Valk, H. W., & Visser, G. H. (2005). Validation of the Continuous
Glucose Monitoring System (CGMS) by the use of two CGMS simultaneously in
pregnant women with type 1 diabetes mellitus. Diabetes Technology & Therapeutics,
7, 699-706; discussion 707-699.
Klonoff, D. C. (2005b). Continuous glucose monitoring: roadmap for 21st century
diabetes therapy. Diabetes Care, 28, 1231-1239.
Klonoff, D. (2005a). A review of continuous glucose monitoring technology. Diabetes
Technology & Therapeutics, 7, 770-775.
Kovatchev, B., Goder-Frederick, L., Cox, D., & Clarke, W. (2004). Evaluating the
accuracy of continuous glucose-monitoring sensors: continuous glucose-error grid
analysis illustrated by TheraSense Freestyle Navigator data. Diabetes Care, 27, 19221928.
Kulcu, E., Tamada, J., Reach, G., Potts, R., & Lesho, M. (2003).Physiological
differences between interstitial glucose and blood glucose measured in human
subjects. Diabetes Care, 26, 2405-2409.
Ludvigsson, J., & Hanas, R. (2003). Continuous subcutaneous glucose monitoring
improved metabolic control in pediatric patients with type 1 diabetes: a controlled
crossover study. Pediatrics, 111, 933-938.
Marangos, P. J. & Papatheofanis, F. J. (2005). Cost-effectiveness of a continuous
glucose monitoring system with trending and rate information to predict
hypoglycaemia and hyperglycaemia in pregnant women with insulin-dependent
diabetes, Journal of Medical Economics, 8, 169-196.
McGahan, L. (2002). Continuous glucose monitoring in the management of diabetes
mellitus. Issues in Emerging Health Technologies, 32, 1-4.
Ministry of Health (2004). An indication of New Zealanders' health 2004.
Wellington: Ministry of Health.
Ministry of Health (2004). Public hospital patient statistics. Available from
http://www.nzhis.govt.nz/stats/hospstats.html Accessed on 07.03.06.
Ministry of Health (2005). Mortality and demographic data 2001. Wellington:
Ministry of Health.
Phillips, B., Ball, C., Sackett, D., Badenoch, D., Straus S,, Haynes, B., Dawes, M.
(2001). Oxford Centre for Evidence-Based Medicine Levels of Evidence. Oxford:
Centre for Evidence-Based Medicine. Available at:
http://www.cebm.net/levels_of_evidence.asp
44
Continuous glucose monitoring devices
Sieg, A., Guy, R. H., & Delgado-Charro, M. B. (2005). Noninvasive and minimally
invasive methods for transdermal glucose monitoring. Diabetes Technology &
Therapeutics, 7, 174-197.
Statistics New Zealand (2005). Demographic trends 2004. Wellington: Statistics
New Zealand.
Tanenberg, R., Bode, B., Lane, W., Levetan, C., Mestman, J., Harmel, A. P., Tobian,
J., et al. (2004). Use of the Continuous Glucose Monitoring System to guide therapy
in patients with insulin-treated diabetes: a randomized controlled trial. Mayo Clinic
Proceedings, 79, 1521-1526.
Tansey, M. J., Beck, R. W., Buckingham, B. A., Mauras, N., Fiallo-Scharer, R., Xing,
D., Killman, C., et al. (2005). Accuracy of the modified Continuous Glucose
Monitoring System (CGMS) sensor in an outpatient setting: results from a diabetes
research in children network (DirecNet) study. Diabetes Technology & Therapeutics,
7, 109-114.
The Diabetes Control and Complications Trial Research Group (1993). The effect of
intensive treatment of diabetes on the development and progression of long-term
complications in insulin-dependent diabetes mellitus. New England Journal of
Medicine, 329, 977-986.
The Diabetes Research in Children Network Study Group (2006). Psychological
aspects of continuous glucose monitoring in pediatric type 1 diabetes. Pediatric
Diabetes, 7, 32-38.
UK Prospective Diabetes Study Group (1998). Intensive blood glucose control with
sulfonylureas or insulin compared with conventional treatment and risk of
complications in patients with type 2 diabetes. Lancet, 352, 837-853.
Wainstein, J., Metzger, M., Boaz, M., Minuchin, O., Cohen, Y., Yaffe, A.,
Yerushalmy, Y., et al. (2005). Insulin pump therapy vs. multiple daily injections in
obese Type 2 diabetic patients. Diabetic Medicine, 22, 1037-1046.
Continuous glucose monitoring devices
45